search this blog

Saturday, February 23, 2019

Catacomb > Armenia_MLBA

It's now clear, thanks to ancient DNA, that Transcaucasia and surrounds were affected by multiple, and at times significant, population movements from Eastern Europe during the Chalcolithic and Bronze Age periods. Based on the ancient samples from what is now Armenia, I'd say that this process peaked during the Middle Bronze Age. But who exactly were the people who perhaps swarmed south of the Caucasus at this time?

The most likely suspects are the various groups that occupied the southernmost parts of the Pontic-Caspian steppe throughout the Bronze Age. They were associated with the so called Catacomb, Kubano-Tersk and Yamnaya archeological cultures. Below is a Principal Component Analysis (PCA) that compares samples from these cultures with those from Middle to Late Bronze Age Armenia (labeled Armenia_MLBA). The relevant datasheet is available here.

Note that Armenia_MLBA forms a cline that appears to be stretching out towards the Catacomb, Kubano-Tersk, Yamnaya and other Bronze Age steppe groups, and this suggests that it harbors significant and probably recent steppe-related ancestry. But PCA plots based on just two dimensions of genetic variation can be misleading at times, so let's check this out with some formal mixture models using qpAdm.

Catacomb 0.234±0.028
Kura-Araxes_Kaps 0.766±0.028
chisq 10.723
tail prob 0.826248
Full output

Kubano-Tersk 0.254±0.030
Kura-Araxes_Kaps 0.746±0.030
chisq 13.535
tail prob 0.633284
Full output

Kura-Araxes_Kaps 0.768±0.028
Yamnaya_Kalmykia 0.232±0.028
chisq 14.454
tail prob 0.564954
Full output

Kura-Araxes_Kaps 0.762±0.029
Yamnaya_Caucasus 0.238±0.029
chisq 15.916
tail prob 0.458816
Full output

All of these models are statistically very sound, and even though I ranked the results by "tail prob", there's nothing in the output that clearly points to any one of the southern steppe groups as the obvious source of the steppe-related ancestry in Armenia_MLBA. But, interestingly, Catacomb tops the ranking, and it probably also makes the most sense based simply on Carbon-14 chronology. So, for now, I'm going with Catacomb.

I didn't get a chance yet to investigate this issue in detail with the Global25. Does it contradict the results from my PCA and qpAdm analyses? If anyone reading this would like to take a close look that'd be great. Feel free to post your findings in the comments below. And if the answer is indeed Catacomb, then what language did these Catacomb-derived migrants, or perhaps invaders, speak? If not proto-Armenian then what?

By the way, please be aware that the Kubano-Tersk samples in my analyses are the same individuals as those featured in Wang et al. 2019 under the label "North Caucasus".

See also...

Early chariot drivers of Transcaucasia came from...

Likely Yamnaya incursion(s) into Northwestern Iran

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...


Ric Hern said...

Looks almost like a Pincer grip around the Black Sea...

Open Genomes said...

Armenia Middle-Late Bronze Age on the Global25 Wards's distance-squared clustering tree

The complete Global25 Ward's distance-squared clustering tree

First, the Bronze Age Global25 restricted nMonte analysis:

Bronze Age and earlier ancestry composition of sample: I1656 Population: Armenia_MLBA Bronze Age Caucasus

While it's clear that Armenia Middle-Late Bronze Age doesn't actually have Beaker Britain ancestry, what we see is something like European steppe-admixed ancestry. If we add up the total steppe components, we get 23.8% Steppe-related ancestry. Common to other individuals from the region from the Bronze Age, we also have 13.2% BMAC/Iran-related ancestry. The largest single source is Kura-Araxes at 37.8% with Darkveti-Meshoko at 9.2%, and an additional CHG at 4.0% which gives a total of 51% native Caucasus ancestry. Kubano-Tersk itself only shows up at 1.2%.

Now for the Chalcolithic and earlier ancestry:
Chalcolithic and earlier ancestry composition of sample: I1656 Population: Armenia_MLBA Bronze Age Caucasus

The native Caucasus-related ancestry goes up to 61.4%, and BMAC/Iran-related ancestry (Tepe Hissar) goes up to 11.0%. The European Neolithic is 14%, and interestingly, we see EHG-related Nordic Late Neolithic at 6.4%. Steppe-related Vonyuchka Eneolithic by itself is just 1.4%, so this won't be the direct source.
The West Siberia Neolithic at 1.6% would be related to various steppe groups.
We have a small amount of Levant Neolithic at 1.4%.

No doubt the combination of European Neolithic mixed with EHG-related Nordic Late Neolithic, Vonyuchka Caucasus Foothills Steppe Eneolithic and extra CHG, combined with a small amount of West Siberia Neolthic, is what looks like various steppe-related ancestries. This is an improbable combination on its own, so it probably does relate to later Bronze Age steppe groups.

The Iran-Neolithic-related ancestry is found in other Caucasus and Near Eastern individuals at this time. This needs an explanation.

The Neolithic and earlier ancestry composition of sample: I1656 Population: Armenia_MLBA Bronze Age Caucasus

Here CHG rises to 31.6%, and European/Anatolian Neolithic-related ancestry to 42.6%. This includes the Anatolian Neolithic ancestry in Kura-Araxes. Some of this may be directly related to Europe, and much of it is certainly directly from the local Anatolian Neolithic.
Iran Neolithic ancestry is now 15.2%. Extra "eastern" ANE + West Siberian Neolithic is 5.8%, and Levant Neolithic is 4.8%. There's no obvious EHG ancestry at this level, unless this is found within I0059 BenzigerodeHeimburg Late Neolithic.

Based on this, I would say that the Steppe-related ancestry in I1656 Armenia Middle-Late Bronze Age is from Bronze Age steppe groups, rather than Steppe Eneolithic groups, including those just north of the Caucasus, with the exception of a small amount of Vonyuchka Eneolithic, which could have been picked up on the way south.

I1656 is a bit divergent from RISE423, who clusters instead with Armenia Chalcolithic.

Lee said...

Just curious if directions of ancestry may be reversed. Or we are looking at common ancestry flowing into both Armenia and steppe groups?

The pcA does NOT support a very strong relationship in any way.

It is relatively certain, in ancient dna, we had a chi high ehg low egh group moving around to the east and west during the late neolithic. Find it in south west Asia to Italy to the Levant. Find it hard to believe this population would not move north as well.
It had less genetic impact than steppe groups to the west, but this could be due to lack of disease vectors depopulating areas.

EastPole said...

I had problems with opening file PCA_of_ancient_West_Eurasia23022019.dat in my PAST 3.22 program.
There was unnecessary ‘enter' in the row 2726:

I removed that ‘enter’ and it works OK:

Maybe others also experienced this problem.

Open Genomes said...

RISE423 Armenia Middle-Late Bronze Age however shows a completely different picture:

Bronze Age and earlier ancestry composition of sample: RISE423 Population: Armenia_MLBA Bronze Age Caucasus

Here Hittite Era Anatolia Middle-Late Bronze Age MA2205 is 33.0%. Afanasievo is 28.6%. Iran-related Tepe Hissar is 10.8%, and Levant Bronze Age North (the Amorites from Sidon) are 8.6%.

Among the actual Caucasian groups, Kura-Araxes is only 8.4%, the female Dzhakutan 2 who clusters with Maykop (a Maykop woman in BMAC) is 6.0%, and Maykop Novosvobodnaya is 2.0%, for a Caucasus Maykop total of 8.0%. Lepinski Vir Anatolian-Neolithic-related ancestry is just 2.O%, and West Siberian / Botai Neolithic is just 0.6%.

This is a very big difference.
First we have to emphasize that population averages generally don't work. These two individuals don't cluster together, because of their very different ancestral mix.

This Middle-Late Bronze Age Armenia individual is a to a great degree a combination of Hittite Era Anatolian and Afanasievo. A case could be made that he's an actual Anatolian speaker.
He barely has any European Neolithic ancestry, so this can't be directly from a late Western Pontic-Caspian Steppe source.

Chalcolithic and earlier ancestry composition of sample: RISE423 Population: Armenia_MLBA Bronze Age Caucasus

Here Kura-Araxes is 43.6%, but Progress and Vonyuchka Eneolithic is 26.2%. His total European-Anatolian Neolithic is 12.6%. Levant Chalcolithic is 14.8%. Extra CHG is just 1.0%. Tepe Hissar Iran-related ancestry is just 1.0%, and West Siberia / Botai Neolithic is just 0.8%

The additional Anatolian-European Neolthic would seem to derive from Progress-Vonyuchka Caucasus Steppe Piedmont Eneolithic, and the same is true for the additional Iran-related Tepe Hissar ancestry. There's almost no additional CHG above these Caucasus Steppe and Kura-Araxes groups. It would seem that the additional ANE / Siberian Neolithic also derives from Progress-Vonyuchka Eneolithic. There is also a higher Levant Chalcolithic, which may come from a movement of Levantines northward during the Early Bronze Age.

Neolithic and earlier ancestry composition of sample: RISE423 Population: Armenia_MLBA Bronze Age Caucasus

The pure CHG component here is 25.8%. The European-Anatolian Neolithic component is 46.8%. The Iran-related Neolithic component is 15.2%, the Natufian is 6.4%, and West Siberian Neolithic / Botai is 5.8%.

The largest single component is European-Anatolian Neolthic. This may be mostly from the residual Antolian Neolithic ancestry in the southern Caucasus region, mixed with some European Neolithic ancestry from Progress-Vonyuchka. The Natufian likewise is related to the northern movement of the Levantine Chalcolithic-related people. Iran-related ancestry is about the same, and comes both from Kura-Araxes and Progress-Vonyuchka. The West Siberian / Botai ancestry comes from Progress-Vonyuchka Eneolithic.

It seems that Progress is the mediator here of Steppe ancestry moving southward. My guess is that this has something to do with Proto-Anatolian, rather than Proto-Armenian.

The question still arises, where did the Iran-related ancestry come from that's lacking in Khvalynsk and Dereivka, but present in Progress-Vonyuchka?

a said...

BU2001.A0101 Beliy Ugol --
Nice work, you managed to cluster Armenian/Catacombe/Sarmatian-Alan! - How did you know?
BU2001.A0101 Beliy Ugol & Kubano-Tersk cluster.

As I posted before BU2001.A0101 Beliy Ugol is a stratified burial--"East-Euro 21,14% North-Sea_Germanic 19,42% North-Caucasian 18,7% Celtic 16,39% Scando-Germanic 16,29% Pamirian 6,91% Volgan 1,13% Baltic 0,02%"

JuanRivera said...

Neither Meshoko nor Maykop contributed to Yamnaya. Instead, the CHG ancestry in Yamnaya is due to a group that stayed since the latest Paleolithic in the North Caucasus Piedmont. Such date is supported by nMonte models largely choosing AG3 over West_Siberia_N as the source of extra ANE. Also, while the mediterranean CHG had impact all the way west to Spain, it had less impact than steppe (in the "normal" Beakers, Beakers_no_steppe are another matter).

JuanRivera said...

There is some common ancestry due to CHG, but it is last seen in the latest Paleolithic. South of the Caucasus, CHG admixed with ANF and Iran_N, whereas north of the Caucasus there was a complex network from Ukraine up to the Altai whose result was EHG and extra AG3-like ANE (different from the MA1-like ANE ancestry already present in CHG) admixtures in steppe CHG, some CHG ancestry in EHG and Ukraine_HG, and after some time, some West_Siberia_N admixture into maybe Piedmont and surely Khvalynsk.

Open Genomes said...

MA2205, an important component of Armenia Middle-Late Bronze RISE423's ancestry, shows no sign of any steppe ancestry.
He's basically a combination of Anatolian Neolithic, Levant Chalcolithic, and Kura-Araxes / Darkveti Meshkoko Caucasus ancestry, with some additional Iranian ancestry:

Bronze Age and earlier ancestry composition of sample: MA2205 Population: Anatolia_MLBA Bronze Age Anatolia

Chalcolithic and earlier ancestry composition of sample: MA2205 Population: Anatolia_MLBA Bronze Age Anatolia

Neolithic ancestry composition of sample: MA2205 Population: Anatolia_MLBA Bronze Age Anatolia

He's a nice mix of 62.4% Anatolian Neolithic-related ancestry, 14.6% CHG, 13.8% Iran Neolithic ancestry, and 9.2% Levant Neolithic ancestry. A perfect example of the "mixing of ancestral populations" in the Near East that took place by the Bronze Age.

Give this, there's a very good case to be made that RISE423 is in fact a Proto-Anatolian, because he's just the sort of mix of pre-Indo European Anatolian (Hattic or Assyrian?) together with a very early kind of Steppe ancestry.

a said...

Rise 397 Kapan LBA,Armenia might be of interest.
PH-4902 quite a bit downstream R-Z2106+

Open Genomes said...

@a I can't find RISE397 in Global25. Are you sure it's there?

a said...

a said...
Open Genomes said...
"@a I can't find RISE397 in Global25. Are you sure it's there?"

Your right, btw RISE397 Armenia_LBA/IA (R1b-Z2106>BY3296)[1048-855 BC] is contemporary to F38
Iran_IA (R1b-L584>Y23838)[971-832 BC], RISE397 shows some steppe ancestry that pulls him towards north Caucasians in Eurogenes PCA. F38 was also found near the site where the undated R1b-Z2103+ sample at Hajji Firuz Tepe. It's an archaeological site located in West Azarbaijan province in north-western Iran. The site was excavated between 1958 and 1968.

Davidski said...

@Lee Albee

Just curious if directions of ancestry may be reversed. Or we are looking at common ancestry flowing into both Armenia and steppe groups?


Yamnaya, Catacomb etc. are entirely of Eneolithic steppe origin. See here...

On Maykop ancestry in Yamnaya

On the other hand, Armenia_MLBA must have steppe ancestry, because it shows R1b-M269, which is not native to the Near East or Caucasus. See here...

Genetic borders are usually linguistic borders too

Drago said...

@ Davidski
The Anatolia MBLA seem to have some structure. Armenia MLBA is not required in any of them.

Anatolia_EBA 60.3%
Armenia_EBA 16.25%
Levant_BA_North 15.45%
Balkans_ChL 8%
Armenian_MLBA 0%
D= 2.9091%

Quite simply, the culture- demography of Anatolian Bronze Age was shaped by Majkop/ Armenia EBA-like groups arriving as early as 3800 BC (Barcin Chalcolithic), seen by the appearnce of Arenical Bronzes. The rest of the period was local adjustments & ongoing contacts, carrying further on with teh K-A phenomenon.

The outlier seems to be MA2203

Anatolia_EBA 51.35%
Balkans_ChL 35.35%
Gonur1_BA 11.85%
d; 2.461%

Also had a look at Balkans with the new data. The much anticipated role for Catacomb migrations is a no-go. Instead, there is an earlier stratum of Majkop-Ukraine Eneolithic; following by integration of these earlier groups with east-Yamnaya (e.g. the basis for Yamnaya Bulg.); then adsoprtion of central Europe MLBA by later BA groups (.e.g the Croatian MLBA

Andrzejewski said...

Whose stupid idea was it that the Anatolian languages date back 6,000 years in situ?

Davidski said...


Both Anatolia_EBA and Anatolia_MLBA have very complex ancestry.

Anatolia_MLBA is largely derived from Anatolia_EBA, but it has extra ancestry from both the east and west, and from the steppe as well.

That's why it can be modeled so successfully with Armenia_EBA and Peloponnese_N as reference pops in qpAdm, no matter which outgroups are used.

Drago said...

I’m quite aware of these complexities
However your model isn’t substantiated . You seem to be searching for something that simply isn’t there; at least not in any substantive, direct or relevant way

Davidski said...


I’m quite aware of these complexities. However your model isn’t substantiated.

I strongly disagree.

Statistically my model looks very good, and, importantly, it also reflects historical and archeological data suggesting that Bronze Age Anatolia was a dynamic and diverse place, with numerous population movements into the region from the west and east.

Davidski said...


Every mainstream publication about Hittite era Anatolia says pretty much the same thing. Here's a recent article that summarizes the mainstream view in its intro.

The Hittites and their Geography: Problems of Hittite Historical Geography

It seems that you want me to accept an alternative narrative, but you haven't outlined what that narrative is exactly, or even roughly.

I hope you're not suggesting that we all go back to Renfrew's kooky model of Neolithic Anatolia as the Proto-Indo-European homeland.

Davidski said...


Oops. I actually meant to say Armenia_MBA not Armenia_EBA in my comment above.

These are the samples that I used in my Anatolia_MLBA (Hittite era Anatolia) model.




And I probably should have emphasized the fact that substituting Anatolia_EBA_Ovaoren with the higher quality Anatolia_EBA_Isparta doesn't change anything in practice.

All you have to do is to drop Peloponnese_N, and you get essentially the same result.

The reason it's necessary to drop Peloponnese_N is probably because Anatolia_EBA_Isparta already has that type of ancestry, since it comes from a site west of Ovaoren.

So everything makes perfect sense.

Drago said...

Right, that article states
“It assumed that the Hittites migratedto Anatolia in the late third to early second millennium”

This just a political history dealing with the Hittite kingdom after its establishment
It provides no facts or data about any of the said migrations. It merely parrots some anecdotes

Davidski said...


It's a brief summary of the status quo in regards to Hittite origins.

The point wasn't to link to an exhaustive discussion on the topic, but to show what the mainstream view was and that my mixture model was very much in line with this mainstream narrative.

If you have an alternative theory and complementary ancient DNA mixture models that rise above this so called mediocrity then feel free to share them.

But don't expect me to suddenly ignore a long-standing academic consensus unless you have something really special.

Drago said...

And you should realise that your post-Kura Araxes model lands square in the Land of the Hatti

Davidski said...


Yes, but note that Armenia_MBA has obvious steppe ancestry.

Drago said...

@ Davidski
“It's a brief summary of the status quo in regards to Hittite origins.”

Well it doesn’t even manage that. It makes an unreferenced statement. It does not mention the views of Darden or Mellart who suggest a much older date
In dealing with prehistory, one needs to look at multidisciplinary scientific data & engage thorough analysis of Sociolinguistics.
For example , you might help answer - is there evidence of mass migrations and when ? is there evidence of elite conquest and by whom ?

As I alluded, the only clear evidence for an elite conquest seems to arrive from the east, into central Anatolia, just prior the pre-Hattic era
. Curiously enough, CENTRAL ANATOLIAN LANGUAGES AND LANGUAGE COMMUNITIES IN THE COLONY PERIOD : A LUWIAN-HATTIAN SYMBIOSIS AND THE INDEPENDENT HITTITES by Petra M. Goedegebuure- paints the very scenario via a thorough analysis of linguistic strata. In other words, the shift you think you’re detecting at the turn of the EBA-MBA might herald the arrival of a non-IE, Hattic group

Davidski said...


My models don't prove or even assume which direction steppe-related ancestry arrived from in Middle Bronze Age Central Anatolia, just that it arrived there during or shortly after the Early Bronze Age.

They're just two-way and three-way mixture models and thus simple reflections of reality, which is probably a lot more complex.

But surely the consensus is that the Hattians were native to Central Anatolia and that the Hittites and Hurrians were the newcomers. So, in fact, the non-Anatolia_EBA ancestry in my models can't be reliably linked to Hattians, but rather to the Hittites and/or Hurrians.

Wikipedia: Hattians

Wikipedia: Hurrians in Anatolia

Halfalp said...

About Anatolians IE origins, cannot it be possible that Late Kura-Araxes experienced a shift towards a newcoming Northern origin without clear cultural change, in other words, a newcomer elite keeping the local culture ( because of economical or cultural prestige? )? One early KA sample was y-dna G2b. Nothing fancy here, something that we could expect from a broadly Middle-Eastern / Iran culture. Then 1000 years later, still Kura-Araxes but we see a sample that is likely R1b-V1636, the same lineage we see from the Southern Steppe at the time of early KA. Why are we actually searching for something very specific, there is no need for a particular intrusive culture into South Caucasus or Anatolia to explain Anatolian Languages, you just need to found the link between Steppe and Middle-East. A little bit more of late KA samples could give an interesting perspective.

Davidski said...


This post isn't about Armenia_EBA (Kura-Araxes), but about Armenia_MLBA, dated to a later period and with very obvious steppe ancestry.

Also, the Armenian reference sample that I use for Anatolia_MLBA is Armenia_MBA, which includes four of the oldest Armenia_MLBA individuals.

Andrzejewski said...

@Davidski "But surely the consensus is that the Hattians were native to Central Anatolia and that the Hittites and Hurrians were the newcomers. So, in fact, the non-Anatolia_EBA ancestry in my models can't be reliably linked to Hattians, but rather to the Hittites and/or Hurrians."

Aren't Hatti and Hurrians both mostly Anatolia Farmers + little bit of CHG admixture?

JuanRivera said...

Did models of Armenian samples months ago, also found steppe ancestry. Unrelated, I think there were three waves of ANE to Eastern Europe. The first was a MA1-related lineage encountering a Dzudzuana-related group in what would become the steppe, giving rise to CHG and Hotu_HG, the second was an AG3-like lineage that gave rise to the overwhelming majority of EHG's ANE, plus the extra ANE outside of both EHG and CHG in Piedmont. The third was one that gave rise to West_Siberia_N and which provided the extra ANE absent in Sidelkino but present in later EHG.

Davidski said...


Aren't Hatti and Hurrians both mostly Anatolia Farmers + little bit of CHG admixture?

I'm pretty sure that they both had quite a bit of CHG admixture, because Hattian and Hurrian are unlikely to have been the languages of Central Anatolian farmers, but rather of migrants from the east.

But my point was that Hattians were native to Central Anatolia, or at least they had a much longer presence there than Hurrians.

In fact, I think it's likely that the Anatolia_EBA samples were Hattic speakers, and they do show significant CHG.

Drago said...

@ Davidski

“But surely the consensus is that the Hattians were native to Central Anatolia”

I wouldn’t rely on the assumptions of text-studying Hittotologist. The arrival of IE groups is beyond their scope
It’s rather difficult to claim that Hattians were “native”; in the sense deriving from Neolithic; when central Anatolia experienced a population replacement after 5000 BC

Davidski said...


There's definitely a consensus that Hattians had a longer presence in Central Anatolia than Hittites (and Hurrians).

This is the mainstream view and it's very simple to verify that it is the mainstream view, so I won't be spending any more time discussing it.

Andrzejewski said...

@Dragos @Davidski I agree with Dragos that Anatolia experienced a population replacement after 5000 BC, more likely after 4000 BC. It's associated with a CHG-rich population replacing the native Anatolian, but most of its ancestry was Anatolia Farmers. This must be the same population that brought over the Minoan Culture.

Would you classify Hatti as a CHG language?

Leron said...

Anatolia is a big place, and it's not profitable to make sweeping generalizations without first understanding the main geographic and cultural divisions that have existed as far as recorded history. We are most familiar with the Hittites, and while their empire stretched a large area of the peninsula they are locked in Central Anatolia, between the Kızılırmak (Halys) and Euphrates rivers. This region was anciently known as the land of Xatti (in Hattian), Sura (Luwian) and later Katpatuka/Cappadocia (Persian). At the height of their power, the Hittites crossed the Euphrates to conquer northern Syria and a small part of the Hurro-Urartian region (Eastern Anatolia). The ancient (Ionian) Greeks who met the descendants of the Hittites/Luwians called them Syrians. They also realized these Syrians lived not only in Cappadocia but also in Syria (Neo-Hittite states) so they called the northern ones White Syrians while the southern branch just Syrians.

The North Anatolian region can be divided into three main parts. The eastern Pontic area, where incredibly barbarous tribes used to reside that even during Ottoman times were difficult to control. They were probably relics of the Neolithic before the Bronze Age expansions. The Hittites didn't bother them much. Further east were the Colchians and other Kartvelian groups. The north central region was the ancient land of Pala or Paphlagonia, the people here shared kinship with the Hittites and spoke a related language. On the northwest was the ancient region of Bithynia and the people there were non-Anatolian speakers, but said to be more similar to Thracians. The Hittites knew of an aggressive tribe known as Kaskas, and they might be from the same pulse that brought the Phrygians and Bithynians into Anatolia.

You can also divide South Anatolia into 3 segments. The west had the Carians and Lycians, who spoke Luwic languages and were fiercely opposed to the Hittites. The south central region was known as Rugged Cilicia, due to its mountainous nature. The Hittites settled this region and established a secondary kingdom known as Tarhuntassa. To the east was the plain of Adana, also called Flat Cilicia, or Kizzuwatdna by the Hittites. The Hurrian presence was substantial here and persisted until the Iron Age although Luwians became the majority. Ethnic Hurrians from Kizzuwatana made up the last line of the Hittite ruling house. After the collapse of the Hittite empire, the southern coast of Anatolia was colonized by Greeks and other groups from Western Anatolia. A lot of their coastal cities trace their origin to the end of the Trojan War.

Western Anatolia, known as Arzawa and later Lydia, was the only other major region in Anatolia that could contest their power against the Hittites of Central Anatolia. There's some alleged evidence that one of their kingdoms (known as Mira) was responsible for the final destruction of the Hittites. Troy (Wilusa in Hittite) was in Mira's sphere of influence but loyal to the Hittite empire, which could explain why the Trojan kingdom disappeared as well. The language of this region was also Luwian, but there existed another Anatolian yet non-Luwian speaking group that eventually dominated, Lydians and Mysians. Further west was the land known as Phrygia. The Phrygians came to occupy several of the areas left abandoned by the defeated Hittites who lost a great deal of territory.

Keep in mind, which region and type of people is being discussed in regards to Anatolians. They could be Hattians or related groups, Central Anatolian Hittites, Luwians, proto-Phrygians, proto-Lydians, or other groups coming from the Aegean or south Caucasus.

Davidski said...


Would you classify Hatti as a CHG language?

No, of course not, because CHG is a somewhat abstract very ancient genome-wide component not linked in any direct way to language.

The only useful thing we can say in this sort of context is that Hattic was probably originally spoken in Central Anatolia by a people whose ancestors came from the east, perhaps from the Caucasus or surrounds, and this is why they showed elevated CHG-related ancestry compared to the preceding Central Anatolian populations.

Leron said...

Correction: Phrygia is further east of Arzawa/Lydia, not west. The Phrygians supposedly migration from the north (Europe) following the Sakarya river until they settled the region to which they gave their name. This made them get stuck between the post-Mira kingdom/early Lydians and the remnants of the Hittites.

Samuel Andrews said...

@Leron, That's a helpful post. Do you know the ethnic makeup of Anatolia before Turkicization?

Modern Greek Anatolians have no Central Asian ancestry but also appear to have no Balkan ancestry. They seem to descend from non-Greek, pre-Turkic Anatolian. If Anatolia was largely Greek before Turkicization they may not have been very Greek in ancestry.

Davidski said...

@Samuel Andrews

There are some present-day Greek Anatolians in the Global25 datasheet.

They don't come out very Greek.

JuanRivera said...

Here are models of EHG: EHG: (EHG:Sidelkino, 2.9119) (EHG:Sidelkino+WHG, 2.9119, EHG:Sidelkino 100% WHG 0%) (EHG:Sidelkino+CHG, 2.9119, EHG:Sidelkino 100% CHG 0%) (EHG:Sidelkino+AfontovaGora3, 1.9988, EHG:Sidelkino 87.5% AfontovaGora3 12.5%).

Interestingly, I found extra ANE outside of Paleosiberian (represented by USR1) and Shamanka_N ancestries in Northeastern Siberia. That extra ANE is AG3-like, instead of MA1-like, as in USR1 and maybe Shamanka_N. So far, I only checked the Chukchi, but it would not be far off to suggest that other peoples in the region share it. Here are models:

Chukchi: (USR1+Shamanka_N, 10.6191, USR1 50% Shamanka_N 50%) (USR1+Shamanka_N+MA1, 10.566, USR1 45.83% Shamanka_N 50.83% MA1 3.33%) (USR1+Shamanka_N+AfontovaGora3, 10.4941, USR1 44.17% Shamanka_N 51.67% AfontovaGora3 4.17%)

JuanRivera said...

Found the same extra AG3-like ANE in Itelmen: Itelmen: (USR1+Shamanka_N, 11.3862, USR1 28.33% Shamanka_N 71.67%) (USR1+Shamanka_N+AfontovaGora3, 11.3655, USR1 23.33% Shamanka_N 73.33% AfontovaGora3 3.33%).

Leron said...

@Samuel Andrews

Vast swaths of people became Hellenized thanks to Alexander's conquests. However, Greek blood didn't reach much outside the coastal cities accessible by the Aegean. Western Anatolia being the most significant in terms of direct Greek relatedness.

More European leaning groups like Phrygians, Thracians and Galatians made inroads of their own at varying degrees in the more northerly parts. An early split off from them were probably the proto-Armenians, who unlike other IE speakers, picked up more Hittite influence.

The Urartians probably knew of the proto-Armenians as the Halitu tribes, who along with the Kaska and Mushki, probably represented Balkanic groups crossing the mostly sparse lands between the Pontic Greek cities and southern post-Hittite towns.

Southeast Anatolia also came under heavy Semitic influence, starting with the Assyrians and later Arabs. In addition to Medo-Persian and Cimmerian intrusions from the east. Anatolia was a melting pot in the surrounding regions, but the center kept more of it's Bronze Age Luwian/Hittite character despite all these new people coming in. It only only until several more centuries of war between east (Persians/Sassanians) and west (Romans/Byzantines) that the central population collapsed from all the brutality that left a vacuum for later Turks to fill.

Samuel Andrews said...

@Leron. Last question. Do you know the origins of Greek Cypriots & Cretans. They are different from mainland Greeks. Also, Slavic+Greek Cyrpiot/CRetan does not result in mainland Greek. So, I think there's a ancient separation between them and mainland Greeks.

MomOfZoha said...

You seem a good person to ask (I hope): Where do you figure the Isaurians fit into all of this? Their Byzantine emperors were called alternatively "Syrian" or "Armenian", and their personal names were Luwian in an otherwise unclassified language. However, my feeling from the extremely iconoclastic religious attitude of the Isaurian emperors in addition to the "fiercely independent mountaineer" nature of the people themselves, is that they have more affinities to Northern Mesopotamian people who share some pre-Abrahamic cultural complex (pre-Kurds, perhaps)...

Also, since you mention Galatia: Did the Gauls leave any noticeable influence upon Galatia, e.g. in a genetic sense? (Should a Central Anatolian with geographically "Galatian" ancestry suspect any actual Gallic influence when some chromosomes show major affinity to British Isles -- or is there some more obvious candidate for the "British Isles" vector into Central Anatolian Galatia?).

Davidski said...


Also, since you mention Galatia: Did the Gauls leave any noticeable influence upon Galatia, e.g. in a genetic sense?

A closer look at a couple of ancients from Hellenistic Anatolia

Leron said...

@Samuel Andrews

Greek tradition says Crete was taken over by Dorians, who are supposed to be a northern Hellenic branch that replaced the older Mycenaean (Achaean) people who wrote using Linear B script and were much better acquainted with the Hittites. Dorians may have become the ruling group, but there would also be Minoan survivals, Lycians, Hurrians, Phoenicians and others that contributed to the population admixture.

Unlike Crete, Cyprus kept more of the older Achaean Greek strain and native Eteocypriot blood. Ancient Cyprus also had mixed elements, with stronger Hittite/Luwian and Hurrian affinities prior to Greek colonization.


I don't recall Luwian or Luwian-like names in the ruling class existing as late as Byzantine times. Although I'm sure there would exist among some of the common people who did not adopt Greek. Anatolian languages may have been moribund or dead by then, but the Isaurians probably had significant Luwian ancestry. It would just be more diluted in the upper class who were mostly of Greek and/or Persian origin.

Persians absorbed various elements of Assyrian culture and administration. One of them being the continued use of Aramaic script and language in the Near East. This is how Semitic influence would penetrate the "old Hittite areas" by way of Persians bringing it along with them.

Galatians came to rule through elite dominance. The majority of the population were pacified Anatolian natives, that had their towns taken as strongholds for the warlike Galatians who would raid around the area. Whatever genetic contribution they had, it was insignificant because they were not able to thrive and mostly kept to themselves outside their predatory activities.

JuanRivera said...

Contrary to my expectations, Yukagirs (even Yukagir_Tundra) have less extra ANE than Chukotko-Kamchatkan-speaking peoples.

Ric Hern said...

@ JuanRivera


Ric Hern said...

@ Davidski

Does this indirectly point to Hungarian Bell Beakers being Proto-Celtic Speakers ? Or when precisely did they become Celtic speakers I wonder ?

Davidski said...


There are probably many European LNBA populations that can be used as substitutes for Hallstatt in that analysis, so the results don't necessarily mean that Hungarian Bell Beakers were proto-Celtic.

Ric Hern said...

@ Davidski

Does the Hungarian Bell Beaker and Northwest European connection with a spread from the Netherlands towards the Southeast point to a possibility that Celtic or Proto-Celtic maybe spread from the Netherlands to Hungary....? Just a thought...

Ric Hern said...

@ Davidski

Okay. Thanks.

Folker said...

Just some corrections to Leton post:
- Kaskas were non IE, and occupied the Northern part of Anatolia (including Pala),
- Phrygians came in Anatolia around the BA collapse (after the fall of the Hittite Empire),
- exact location of Anatolian languages is much discussed. In any case, about IE languages at the beginning of MBA, Palaic was spoken in Pala, Hittite probably only around Kanesh (as a minority language), Luwian around the Konyan Steppe. Other languages from the Luwian branch (Carian,...) are not attested before IA, and were probably elsewhere during BA (iatus in occupation during MLBA).

Very funny to read somebody citing Mellaart, despite his hoaxes.

Ric Hern said...

@ Davidski

Sorry maybe a stupid question but how close is Hallstatt_Bylany to the Netherland Bell Beakers ?

Aram said...


The last few months I read a lot of archaeological stuff and I came to the same conclusion, that Armenia MBA was affected by Catacomb culture. And there is even an intermediate culture between South Caucasus and Steppe. It is called Prisulak culture (2400BC-1500BC) in the Caspian litoral part of Daguestan.
Unfortunately I couldn't find anything in English. Here it is a summary in Russian about Prisulak culture,

The last paragraph speaks about Catacomb tribes.
The next stage of migration toward South Caucasus is called Early kurgans of Transcaucasia (2400-2200BC). From this culture emerges Trialeti-Vanadzor culture (2200BC-1600BC).

It must be noted that Kubano-Tersk culture also was catacombocized at it's late stages.

Davidski said...


Yeah it is often repeated in mundane works and by certain armchair historians which post here, but if you are going to pretend to discuss the subject, then you have to bring the kind of works I outlined, especially something as murky & complex as this.

Right, complexity and nuances are all good when they contradict the "armchair historian" consensus that you hate and back up your pet theory. Otherwise they're no good.

I've seen this sort of approach before plenty of times. I'm not impressed by it.

How's that theory working out about the Hungarian Bell Beakers? You know, the one in which they all run away in a panic from the Carpathian Basin to Western Europe because they're getting their asses whooped by Nagyrev people or whoever.

Getting widespread academic acceptance is it? Let us know how that goes. And I'll be reminding you about this for a long time.

One would take your highlighting of Armenia MLBA as signifying an eastern entry, no?

No, that's what I tried to explain above.

Armenia_MBA might be a proxy for a couple streams of ancestry that can't be explained in two-way or three-way models, including gene from the steppe via the Balkans.

Aram said...

Now concerning Your Anatolia MLBA models.

The main problem with this models is that there is no archaeological trail from South Caucasus towar Anatolia in MBA. The reason to this is that Kura-Araxian cultural elements continued to persist in East Anatolia. So it was quite problematic to this MBA steppic groups to move further West because they must cross densely populated regions and various mountain ranges in East Anatolia. Notice for example to move from Daguestan to Armenia is not very hard because Azerbaijan is mostly a steppic place like their homeland.

On the other hand it seems there is some not well understood migration from Azerbaijan MBA to NW Iran. You will remember Hajji-Firuz_BA with a bloated level of steppe ancestry.
This sample falls in so called Urmia Ware period that comes after Kura-Araxes in NW Iran. Another proposed name for Urmia ware is the Yayla culture. Yayla means pasture.

Those Urmia ware groups also tried to move further West. But it seems they were not very successful because their further migrations was blocked by so called Khabur and Nuzi ware folks who are associated with Semitic and Hurrian respectively. The Urmia ware in this context falls into Gutian period in NW Iran.

And the last thing we haven't yet see any R1b-Z2103 in Bronze Age Anatolia. While we haven't any problem to find them in Armenia and Iran. This could be a sampling issue but it could mean that there were no much Z2103 in Bronze Age Anatolia.

Drago said...

@ Aram
That’s interesting, because Trialeti type material is also found in Greece (types of swords and spears, I think). Of course, you have the Alacahoyuk shaft graves in north central Anatolia too.. .

Davidski said...


The last few months I read a lot of archaeological stuff and I came to the same conclusion, that Armenia MBA was affected by Catacomb culture. And there is even an intermediate culture between South Caucasus and Steppe. It is called Prisulak culture (2400BC-1500BC) in the Caspian litoral part of Dagestan.

Thanks, this makes sense.

But what were their linguistic affinities?

Drago said...

@ Davidski

Lol. Okay, your CWC-BB model is solid, I already stated that. Btw I wasn’t referring to you.

Folker said...

Difficult to say which uniparental markers related to Steppe ancestry are to be found in Anatolia in relation to Anatolian IE migration.
Could be rare subclade of R1b, or I2a S12195 (found in Yamnaya Bulgaria) or another one.
A consequence of the possible link between CW and BBs, or a side effect, would be that Yamnaya is not the source of most modern IE languages, meaning that Bulgarian Yamnaya could have been at the right place and time to be the ancestors of Anatolian IE. But we will see. Too few samples are published.

Aram said...

Here is a link about the persistence of Kura Araxians in East Anatolia even in 2nd millenium BC
Page 681


Concerning the linguistics.

Well Marija Gimbutas as far as I know placed ancestors of Armenians, Greeks and Albanians into Catacomb culture. So at first glance there shouldn't be any problem with linguistics. The language of Gutians is not known. But some of their names do have IE etymology. And especially their tribal name look very similar to Getae a Thracian tribal name.
In Armenia the kurganic theory is not popular, but some scholars who continue Jahukyan's traditions are making research about steppe origins of Armenians. One of them is the linguist Hrach Martirosyan (

He and a group of other scholars proposed a theory that the Armenian language descends from this MBA groups. Their argumentations are also linguistic.
+Presence of Paleobalkanic names in Hayassa ( like Karannis ). Notice this kingdom formed at 1400 BC which is supposed to be before any Phrygians were present in Anatolia. And Hittite empire was still functioning.
+Presence of Armenian words and toponyms in Urartian cuneiforms.
+Preservation of laryngeals in Armenian, which could be impossible if Armenians stayed very long time in contact with Phrygians and Thracians.
+Presence of genuine Hittite loanwords in Armenian.

They called their theory Dragon stone theory. Because those groups started to erect Stellaes similar to those found in Steppe with one exception. That they depicted Dragons on this Stellaes. This tradition of erecting stellaes persisted quite longtime in Armenia and after the adoption of Christianity it was transformed into Khachkar ( Cross Stone )

On the other side those who don't believe in Steppe theory, propose that there is no a need for imagining migrations. The PIE homeland was in Near East and they even cite Reich and recent Wang paper that proves Near Eastern origins of Armenians. Wang made such a proposal.

This is the theoretical part, and my next post will be about genetics and the possibilites of LBA/Iron Age migrations via Anatolia that are so popular among Western scholars.

Aram said...

OK. Just an addition. Because Armenian is usually linked with Greek, then we need to understand from what culture Greek emerges.. ( Not all scholars believe there is a strong link. For example Clackson from Oxford university dismissed it )

Currently the main issue I see is that we don't have yet good understanding how Greeks,Thracians and Phrygians formed. If he had a better understanding how this nations formed (from what culture) we would have a much better understanding about Armenian.
I will just add that based on genetic data the most realistic theory about links of Armenian and Balkanic nations is the Kortland's theory Armeno-Thracian theory. Thracians are attested (according Greek sources) not only in western Pontic region but also in North of Caucasus and even in Iran. And the term Mush-ki is a bvious parallel with Moesan in Bulgaria. We have also Trerk that You can compare with Treri.

Kortland proposed a division were Greek clusters with Phrygian while Armenian cluster with Thracian. And they form a node on upper level.

Greek and Phrygian are centum while Armenian and Thracian are satemised languages.


There is lineage that fits well for Hittites
It is this branch of R1b-M269
It is not a Z2103 and it is a Eneolithic split. And it is not rare in Turkey. At last not very rare. There is some 2-3% in modern Turkey. While the I2 that You mentioned is very very rare, practically inexistant.

Folker said...

The problem with Y haplogroups modern distribution is that it is not representative of ancient distribution. Without ancient samples, we will not know.

But interesting idea.

Grey said...

*if* early metallurgy was split into two stages

101: simple working of native soft metals found on the surface or from rivers

102: actual mining and smelting of ore from rocks (requiring lots of wood)

then if we assume that not every population living on a source of soft metal ore figured it out at the same time it would follow there was likely movement of miners/artisans from the people(s) who figured it out first into other regions with ore and some of those movements might include moving into remote mountainous regions with hostile natives and without a lot of surplus food available locally so

1) miners/artisans
2) people to provide food to the miners
3) people to provide wood to the miners
4) people to provide security

with those distinct functions possibly performed by two or more originally distinct ethnic groups e.g miners + herders?

Drago said...

@ Davidski
I’m sure you know this , but of course BB was succeeded & territory taken over by Nagyerv

@ Aram
Interesting theory; but I doubt you’ll be correct
Catacomb is an Azov-Caucasian group. It doesn’t provide fits for Balkan BA; nor is it really present

a said...

Blogger Grey said...
"*if* early metallurgy was split into two stages

101: simple working of native soft metals found on the surface or from rivers

102: actual mining and smelting of ore from rocks (requiring lots of wood)

then if we assume that not every population living on a source of soft metal ore figured it out at the same time it would follow there was likely movement of miners/artisans from the people(s) who figured it out first into other regions with ore and some of those movements might include moving into remote mountainous regions with hostile natives and without a lot of surplus food available locally so

1) miners/artisans
2) people to provide food to the miners
3) people to provide wood to the miners
4) people to provide security

with those distinct functions possibly performed by two or more originally distinct ethnic groups e.g miners + herders?"

Interesting ideas, keep in mind the spheres of influence for copper/bronze smelting. Balkans/Carpithian,Maykop,Uruk,Kargaly

Yamnaya was in contact with 3 of these Zones of metallurgy. Yet they differed. They worked and were mobile on steppe without building/leaving evidence of settlements[Kargaly]only burials.

IO444 Kutuluk-3335-2881BCE pure copper bar celt.burial rite---R1b-Z2109-KMS75
BU2001.A0101 Beliy Ugol-2867-2581 calBCE -- bronze weapon burial rite--R1b-Z2109-Y20993+

Andrzejewski said...

In this article it's said that EHG were the dominant group in Yamnaya, and that CHG constituted only 25%. They mentioned that WHG were another 6% but I'm surprised none was brought up regarding the significant EEF contribution (10%-20%).

Quote: "A little more detail has been figured Yamnayans. They are blend of three ancestral populations. The dominant group are the Mesolithic Eastern Hunter-Gatherer (EHG), made of the Y-haplogroups R1a and R1b. The Caucasian Hunter-Gatherer (CHG) admixture, makes up another 15-25% of their genomes. This admixture is mainly associated with Y-haplogroup J. The third admixture was the Western Hunter-Gatherer (WHG), representing about 6% of Yamna genomes, which composed of haplogroup I, and to a lower extent also C1a2 and F."

Leron said...


The ethnic identity of the Kaska is far from conclusive. In particular I don't think we can write them off as non-IE simply because they appear too amorphous. Almost nothing of Kaska names and words have been preserved. The older Hittologists conjecture them to be related to Circassian or Colchian people and don't bother to look at other connections.

What is more clear about the Kaska is that they were a semi-nomadic population that would encroach on Pala and northern Hatti and then retreat back to their forested lands. They were composed of many tribes, each with its own chief, and when needed formed confederations. They first appear in Hittite written records when discussing events to their northwest. At one point the Kaska overran the Hittite capital and the Hittites had to flee south until it was reconquered by Hattusili III. The king of Arzawa mentions these acts done by the Kaska to the pharaoh of Egypt. After the fall of the Hittites the Kaska, along with the Muski, appear near northern Assyrian territory that was formerly part of Mitanni.

Based on the movements of the Kaska, from the late Bronze Age to the Iron Age, they were migrating west to east. The name Kaska is never seen again yet people don't vanish; they are known by new names in new territories. These Kaska who settled east past the Hittite lands are in my estimation the source of the proto-Armenians. This coincides with the characteristic of Armenian language, showing only slight contribution from Hittite and Hurro-Urartian but a much greater influence from Iranian languages to the point it became satemized. The proto-Armenians were present in eastern Anatolia between the fall of the Hittites and the end of the Urartian kingdom, yet fully spread over Armenia (classical Armenia, not the current republic) only after centuries of Persian rule in the area. In this way, not having being directly conquered by the Achaemenids they were able to preserve their language unlike the Urartians. In comparison, the Iranian speaking Kurds represent in part remnants of the Hurro-Urartian groups that had a complete language replacement impossed on them.

The next question is, is a greater proportion of steppe found in northern Anatolia an indication of Kaska/proto-Armenian populations?

Davidski said...


In this article it's said that EHG were the dominant group in Yamnaya, and that CHG constituted only 25%. They mentioned that WHG were another 6% but I'm surprised none was brought up regarding the significant EEF contribution (10%-20%).

That article is total garbage. Please don't discuss it here.

Samuel Andrews said...

What do people think about the fact Egyptian_mummy_769-560calBCE:JK2911 has roughly 10% Steppe ancestry. Where the did he get it from. Iranians? Hittites?

Davidski said...

@Samuel Andrews

What do people think about the fact Egyptian_mummy_769-560calBCE:JK2911 has roughly 10% Steppe ancestry. Where the did he get it from. Iranians? Hittites?

You should treat that sample as a preliminary version. Hopefully I can get more reliable versions of those ancient Egyptians soon.

Open Genomes said...

The Global25 nMonte analyis of the pre-Hittite Early Bronze Age Anatolians from Ovaroen:

Pre-Bronze Age ancestry composition of sample: MA2212 Population: Anatolia_EBA Bronze Age Anatolia

43.6% Anatolian Neolithic
21.6% Iran Neolithic-related / BMAC ancestry
21.0% Levant Neolithic and Chalcolithic
8.0% CHG
5.2% Kura-Araxes

The most striking thing about MA2212 is the high level of Iran-related ancestry compared to the previous Anatolian Neolithic farmers, who totally lack Iran-related ancestry at all.

Pre-Bronze Age ancestry composition of sample: MA2210 Population: Anatolia_EBA Bronze Age Anatolia

32.0% Anatolian Neolithic
7.2% Ukraine Neolithic outlier / European Neolithic-related ancestry
13.0% Kura-Araxes
9.6% Darkveti-Meshoko
12.8% Levant Chalcolithic
4.2% Hajji Firuz Chalcolithic
7.6% Iran-related ancestry

Here we have an additional source of Early Neolithic ancestry aside from the Anatolian Neolithic, and this is from a source at the eastern edge of the Neolithic expansion in Europe.

There is more Kura-Araxes and much less Iran Neolithic-related ancestry.

Both individuals have a substantial contribution from the Levantine Neolithic-Chalcolithic, from before there was much Anatolian ancestry in the Levant.

Are these Early Bronze Age Anatolians Hattic speakers?

Again, what do they have to do with Iran?

Folker said...

Identifying Kaskas with proto-Armenians is your pet theory, nothing more. The only links I find for Kaskas is with Hattians (as some of their names are looking like old Hattians names and some deities seem similar to the Hattians ones). And Hattic is a non IE language.

Leron said...


Hittites made the distinction that the Kaska worshiped gods different from them. Hittites absorbed Hattian religion and used the same Hattian appellations to refer to distinct deities around Anatolia. Similar to how the Greeks called distinct Anatolian gods as Zeus, Hermes, Hercules and so on. We know absolutely nothing about the Kaska language itself, so it can't be identified with any certainty as belonging to IE or any another language family.

@Open Genomes

The eastern contributions into BA Anatolia look to be from the Kura-Araxes expansions of Hurro-Urartian and related groups. Hurric tribes were also coming from Iran into Mesopotamia, such as the Turruku (perhaps precursors to the Mitanni), supplementing the older Hurrian strata already well established in cities like Urkesh/Tell Mozan and in regions like Cilicia.

Anatolian Neolithic signal is probably that of the Hattians. And in that particular location the samples were from, it was a region of heavy Luwian presence, which may have already absorbed the Hurrian groups that were previously living there prior to Hittite conquest.

Arza said...

Re: qpAdm vs. G25

G25 scaled

qpAdm model:

Kura-Araxes_Kaps 76%
Catacomb 24%
Distance 3.0168%

Probably the best 2-way model in G25 that utilises Kura-Araxes_Kaps:

Kura-Araxes_Kaps 73.5%
Oy_Dzhaylau_MLBA 26.5%
Distance 1.891%

4-way model based on the result from a panel of 239 populations:

Kura-Araxes_Kaps 54%
Yamnaya_Ukraine 19.8%
Levant_BA_North 14.8%
Yamnaya_Bulgaria 11.4%
Distance 1.5324%

With all Kura-Araxes varieties lowest distance is obtained with Yamnaya_Caucasus:

Kura-Araxes_Kalavan 25.2%
Kura-Araxes_Kaps 20.5%
Yamnaya_Bulgaria 17.1%
Yamnaya_Caucasus 14.3%
Levant_BA_North 12.1%
Kura-Araxes_Velikent 9.7%
Kura-Araxes_Talin 1.1%
Distance 1.3988%

Andrzejewski said...

@Leron @OpenGenomes @Folker

Kaskians originated in NorthWestern Anatolia, maybe even came from the Balkan. Does it explain the extra EEF/ANF admixture in the model?

Moreover, they merged with the Proto-Colchis. Does it explain the fact that modern Georgians have 30% Anatolia farmers?

Hard to tell.

Folker said...

"We know absolutely nothing about the Kaska language itself, so it can't be identified with any certainty as belonging to IE or any another language family."

It is not true. We have some names.

It could be interesting that you gave us some sources.

About Kaska:

By the way, you seem a bit confused, as you are mixing different periods of time.


"Kaskians originated in NorthWestern Anatolia, maybe even came from the Balkan. Does it explain the extra EEF/ANF admixture in the model?"

Kaska's origin is probably the same as Hattians, and could be the result of migrations from Caucasus. Their territories are located in the NE of Hatti, and they expanded (XVth century) to Pala (Paphlagonia). So to the West. Because of Kaska, Hittites lost access to the Black Sea.

Arza said...

Re: Ostrogothic off-topic

I've managed to turn admixture graph from this poster:

into CSV file:

which can be used to recreate PCA for West Eurasia:

The PCA above was made using West Eurasian pops and it's a slightly rotated 3D view of PC1, PC2 and PC4 from a four dimensional PCA. Somehow this allowed to recreate the WE PCA despite that the first tree dimensions have troubles with distinguishing modern Europeans from plain WHG-EEF mixtures. Even stranger is fact that despite the problem mentioned above the PCA clearly fleshes out the so-called Balto-Slavic drift and the substructure of European Hunter-Gatherers.

Here is a cross check with the PCA from the poster:


10 closest pops (using ADMIXTURE output):
Yamnaya_Bulgaria 4.57917852801
HungarianScythian3 4.83910969522
Hungarian_Med 6.4191867145
Hungary_BA 9.35492313484
Romanians 10.1207850908
Macedonians 11.2492842575
HungarianScythian2 11.2688259071
Serbians 11.4899865569
Sweden_LNBA 11.5897520511
Bulgarians 11.6263585745

Clearly it has some significant Balkan shift.


GoldenHordeEuro 5.8420080303
Ukrainians 6.49512058499
Scy_Ukr8 7.39425507219
Slovaks 7.52864913911
Swedes 7.65239659036
Russians_South 8.36342824104
Germans 8.88181697817
Orcadians 8.95848777749
Lithuania_BA 9.78962413009
Belarusians 9.93927193401

This sample is a little bit problematic. In the PCA above it basically looks like a pure WHG-EEF mix (Pitted Ware guy from Gotland?!?). Yet distances to moderns are small (PCA provides explanation for this as it lumps moderns with WHG-EEF in the first 3 dimensions - hence the closeness).

So the verdict is that on a proper PCA it will land somewhere between Balto-Slavs, Ukraine_Eneolithic, WHG and GAC. Anyway, it has loads of WHG and as the PCA hints it's rather the classic WHG, not the WHG of the Baltic_BA type. Strange.


Scy_Ukr9 2.095717158
Sweden_LNBA 2.3112287431
Northern_LNBA 2.42076064484
Bell_Beaker 3.27910824776
Central_LNBA 4.90049133796
HungarianScythian5 5.099066653
Petrovka 5.73669097184
HungarianScythian4 6.92727790484
HungarianScythian2 7.30638852745
Poprad 7.3733203016

Sweden_LNBA? Northern_LNBA? A Goth? Unless Y-DNA will say "yes", not really. Rather it's someone from a local subpopulation without (or with very little) East Asian admixture. Is such scenario possible? Scy_Ukr9 on top says or even screams "yes".

Scy_Ukr9 for comparison:

Sweden_LNBA 0.596803957662
Chern3 2.095717158
Northern_LNBA 4.08334358061
Central_LNBA 4.38853397437
Petrovka 4.50315141723
Bell_Beaker 5.07352409507
Maitan_MLBA_Alakul 5.70994024604
Sintashta 5.87363229516
Kairan_MLBA 6.0294539309
Sintashta_MLBA 6.56967357775

This Scythian is also interesting:


GoldenHordeEuro 3.86865048662
Ukrainians 5.6759084583
Ukraine_Eneolithic 6.41827160144
Slovaks 6.57354839442
Chern2 7.39425507219
Russians_South 7.81610499256
Germans 7.97284195281
Swedes 7.98465110219
Central_LNBA 8.19896985544
Mordovians 8.21173762423

Romulus said...

Are these Chernyakhov samples the same ones confirmed for I1 already, or new ones?

Lee said...


I get a fairly good mix for Beakers Netherlands as a 3 way mix (via qpAdm)

of Latvia-HG, CWC-Germany and LBK_EN
fixed pat wt dof chisq tail prob
000 0 9 6.260 0.713639 0.145 0.345 0.511

Not as a "pure" corded ware population

as less favorable match but with better Error would be a Iberia BB instead of LBK-EN

fixed pat wt dof chisq tail prob
000 0 9 10.168 0.337081 0.085 0.560 0.356

I can even get a decent model of just a two way of Iberia BB and CWC germany
fixed pat wt dof chisq tail prob
00 0 10 11.754 0.301861 0.552 0.448

So BB-neatherlands definitely has Corded ware in it. But I am getting about a 50/50 mix of Corded ware and essentially late neolithic. The neolithic could be local (probably) or Iberian bell beaker.

For Beaker_Central Europe

Iberia BB is NOT a fit
but Yamnaya_Kalmykia and LBK_EN give a pretty decent fit
fixed pat wt dof chisq tail prob
00 0 10 6.366 0.783652 0.276 0.724

Substituting in Sintashta_MBA for Yamnaya gives a better fit but higher error

fixed pat wt dof chisq tail prob
00 0 10 2.514 0.99067 0.722 0.278

If I split BB_Iberia into two groups with BB_IberiaS containing I0461
I5665, I6466, I6467, I6471, I6539, I6623, I6472

BB_IberiaS--does indeed come up with a mix of CWC-Germany and Iberia_Chalcolithic that is pretty convincing

fixed pat wt dof chisq tail prob
00 0 11 4.829 0.939174 0.621 0.379
and lowish error: std. errors: 0.026 0.026

So I can see the mix of Central europe into Iberia with these samples. Maybe second generation mixing?

Ric Hern said...

@ Lee Albee

How close is Hallstatt_Bylany samples to Beaker Netherlands compared to Hungary Bell Beakers ?

Davidski said...

@Lee Albee

I can assure you that Beaker_The_Netherlands doesn't have any Latvia_HG ancestry.

It's actually just a more westerly version of Corded Ware with extra hunter-gatherer-rich farmer ancestry from Northern and probably Atlantic Europe (but with nothing from the Baltics or Iberia).

Beaker_The_Netherlands is a very good fit for the steppe-rich admixture in all other Beakers, so it's likely that the steppe ancestry within the Beaker complex came from the steppe via the Lower Rhine region.

But Iberian Beakers didn't have much of a genetic impact outside of Iberia.

You're getting reasonable fits using these tools, but I have to tell you that your models and conclusions are kind of out there at the moment. That's why I suggested that you should have a go at using the Global25/nMonte. It works very well with ancient data, even as an unsupervised test.

Have you actually read the Olalde et al. Beaker paper from where this data came from? If not, you should read it.

Open Genomes said...

@David @Leron

Anatolian Neolithic Farmer ancestry dominates in many regions even until the Late Bronze Age.

What could the linguistic affiliations of these Early Neolithic Farmers have been?

Remember, a related language would also have to be a substrate to the later Indo-European languages of Europe.

Some candidates, from west to east:

1. A Tyrsenian language related to Etruscan - but was one like Lemnian present in pre-Iron Age Anatolia?

2. The Linear A / Minoan language (which may be affiliated with some Near Eastern language, or not.)

3. Hattic (including Kaskian) - was this the language of the early natives of Anatolia, or a later, Caucasian language related to Northwest Caucasian? (Hattic doesn't seem to be related to Etruscan at all.)

4. Kartvelian - although there's no evidence for Kartvelian before the Hellenistic Era, and Kartvelian shares some similarities (e.g pronouns) with Proto-Indo-European.

5. Hurro-Urartean - Hurrian has the advantage of being present across most of the early Neolithic region of the Norther Fertile Crescent and also in Central Anatolia. Like Etruscan and Minoan, Hurro-Urartean seems to have lacked a strong voiced-voiceless consonantal distinction. Hurro-Urartean may or may not be related to the Northeast Caucasian languages in a greater "Alarodian" language family.

6. Sumerian - it has many terms and distinctions for Neolithic cultural items such as domesticated cattle and bread. There's no evidence however that Sumerian was spoken in the north before the Early Bronze Age Uruk Expansion.

7. Elamite - however, there's no evidence that Elamite was ever spoken west of Iran, and Elamite is more likely to have been the Iranian Neolithic language.

Any opinions?

Leron said...


The paper you linked made no concrete analysis of the Kaska language. It only brought up a very fragmentary list of Kaska gods that were written according to Hittite interpretation, not from actual Kaska people telling of their own gods. The actual names were hidden behind Hittite logographic symbols and the author only guesses stood for names of Hattian gods. A few other names were from gods the Hittites worshiped themselves like Telipinu.

The best answer to the location of the Kaska hinges on the geographic position of a land known as Durmitta (from which the Kaska also inhabited according to Hittite sources). In early Hittitology Durmitta was assumed to be northeast of Hatti, and a lot of works were based on this assumption. However, as more cuneiform records were found and studied Durmitta's location was secured to be west of the main Hittite land. This is further corroborated by the Kaska tribes reaching as far as Nenassa (near Lake Tuz) to the south of the Kızılırmak river. This gives a picture of the Bronze Age Kaska bordering the Hittite core from the west of said river, all the way from Pala in the north to Durmitta in the south.

Bob Floy said...


If EHG does indeed have some BE ancestry, do you think that it could end up eating some of Yamnaya's already small amount of EEF? Could some of that be native EHG basal?

Ric Hern said...

@ Open Genomes

I think if CHG formed near the Lower Don/Steppe Piedmont then maybe the PIE and Kartvelian similarities could predate the formation of both PIE and Proto-Kartvelian. So a deep Mesolithic connection maybe.

Mem said...

At Chalcolithic era,Anatolian Neolithic ancestry decreased to half and CHG increased very high level in Anatolian and Greece.

So,Hattian and Minoan were Caucasian languages.

Kartvelian languages family expended at late bronze/early iron age from kolkhis plain.

Only Tyrsenian and Vasconic families direct came from AN.

Leron said...

@Open Genomes

Kartvelian is probably a surviving branch of the Neolithic Anatolian language. Hattian would be another branch, but due to relatively few remains its more difficult to study. Minoan could be a mix of this and a later CHG/Iran related language arriving during the early or middle Bronze Age. The best evidence pointing to the general direction of Kartvelian is its connection with the pre-Greek substrate language.

Davidski said...

@Bob Floy

If EHG does indeed have some BE ancestry, do you think that it could end up eating some of Yamnaya's already small amount of EEF? Could some of that be native EHG basal?

Not exactly sure what you mean, but EEF ancestry in Yamnaya is an issue independent from EHG, because it still shows up in Yamnaya at around the same level when EHG is used as a reference in mixture models. See here...

Another look at the genetic structure of Yamnaya

Note that Yamnaya can't be modeled as a straight two-way mix between EHG and CHG. The models need EEF plus some extra WHG (or European_MN, which is the same thing) to work.

Davidski said...


How close are Hallstatt_Bylany samples to Beaker Netherlands compared to Hungary Bell Beakers?

Overall, they're about the same. Although it depends what measure is used.

Also, Hungarian Beakers are genetically very heterogeneous, so some are almost identical to the Hallstatt samples while others very different.

Bob Floy said...

"Not exactly sure what you mean"

I was asking if perhaps some of the EEF in Yamnaya was actually basal ancestry from EHG, but...

"Note that Yamnaya can't be modeled as a straight two-way mix between EHG and CHG. The models need EEF plus some extra WHG (or European_MN, which is the same thing) to work."

...answered that question, thanks.

Ric Hern said...

@ Davidski

Thanks. I was wondering about the possible continuity of Netherlands related Beaker ancestry after their spread towards Hungary. So if I understand correctly it looks like there were at least some Genetic continuity in that area from the Bronze Age into the Iron Age which I think is interesting regarding Possible Linguistic continuity or relatedness....

Folker said...

A southern location of Durhumit (as proposed by Forlanini) has not reached consensus. More recent works (Kryszen, de Martino) are backing a location a bit Southern than the traditional one (around Samsun), around Merzifon. So not in the West, but in NE of Hatti.

Bogdan said...

Indo European cultures, languages and significant mutations from early to mid-bronze age mostly derive from the area between Dnieper/Don rivers, in modern day Ukraine.

Aram said...

Thanks. The presence of Yamna Bulgaria in Your models is not surprising. This is because Armenia MLBA is not exactly aligned between Yamna/Catacomb and Kura Araxes. It has an extra ANF/EEF shift. This means that either the group that moved from North had a extra EEF. That must be from Ukraine.
Or the substrate is not exactly Kura Araxes. But something like Hajji Firuz Chl or hypothetic Leyla Tepe. It is possible that some Leyla tepe groups survived in Azerbaijan lowlands.
Or alternatively some independent migration from Anatolia. More sampling is needed to understand the origin of that EEF shift.

Bogdan said...

The entire Dnieper watershed in particular at this time was literally teaming with life and easy food resources, unimaginable by today’s standards, far eclipsing other areas of Europe. Literally and figuratively a “tree of life”, from the numerous headwaters spanning out like a canopy through dense rich forest lands S. Baltic (mostly modern day Belarus) down the Dnieper tree trunk and steppe lands with rich virgin loam soils, onto the bounty of the Black Sea....

Aram said...

About Hittites. Many comentators here assume that the age of split of Proto Anatolians is overestimated because of substrate or something else.
Well let's see how PIE tree behaves in computational calculations that are calibrated using known historic splits. Like the divergence of Romance languages.
In Gray Atkinson model Hittite split at 6700BC. !!!
In the Chang et al. without constraints the Hittite split more than 5500BC.
Only after applying constraints they get that 4200BC
But after the constraints the proto Iranian drop to 2000 ybp and Romance languages divergence start at 1200ybp which offcourse is not realistic.
So the age of 4200bc is not a overestimation. It is a artificial underevaluation to account substrates and to fit to that Eneolithic Steppe timeframe.

Gaska said...

@ Lee Albee

This Iberian mixture is more than evident in the case of mitochondrial haplogroups, not so much in the Netherlands where 98% of the samples studied by Olalde belong to the Bronze Age, not to the Chalcolithic (where therefore there is more influence of the CWC and the German Neolithic cultures). That is the reason why the Dutch BBs are so similar to the British.

There is a lot of archeological, genetic and anthropological evidence that shows the mix of the CWC and the BBC in Central Europe. Trying to deny insistently the Iberian impact does not lead anywhere. Your models, the only thing they do is confirm what is already proven enough, and they are very useful to understand European prehistory.

David Reich, “The Importance of Ancestry Outliers in Large Sample Size #aDNA Studies”. Forthcoming work led by Iñigo Olalde: probable 1st-generation N. African migrant in 3rd millennium BC Iberia-

You know that Olalde discovered a case of Mit Hap L2, typically African in the BB site of Las Yeseras (Madrid). The geneticists try to give importance to their discoveries by shaking public opinion (Cheddar Man, African outlier in Spain, African mix in Sicily, African male haplogroups in the Balkans and Greece etc ..) with the theme of Asian and African migrations in Europe. I do not know what your intentions are, but I think you should focus on solving the problems that they have created before looking for more controversial issues. On the other hand, the contacts of Africa with the south of Europe are documented in mesolithic deposits of Portugal and Spain (cave of Nerja). It is normal as soon as it takes a few hours to cross the strait, the stranger would be a total isolation of Europe.

Gaska said...

obviusly I wanted to say " I don't know what their intentions are"

Aram said...

In this way, not having being directly conquered by the Achaemenids they were able to preserve their language unlike the Urartians. "

Achemenid empire conquered all Armenia both west and east. So You must find a more realistic explanation why Armenian survived and Urartian didn't.


Aram said...


And what You think about Multi Cordoned Ware. Didit have an impact on Balkanes?

Gaska said...

@ Open Genomes said..."Anatolian Neolithic Farmer ancestry dominates in many regions even until the Late Bronze Age". What could the linguistic affiliations of these Early Neolithic Farmers have been?

Etruscan, Sardinian, Basque

Drago said...

@ Arame

And what You think about Multi Cordoned Ware. Didit have an impact on Balkanes?''

Maybe indirectly, insofar as individuals & cultures were always in contact, but nothing marked, that stands out.
This MCW occurs after Catacomb and before Srubnaya, and like its predecessor, is found predominantly east of the Dnieper.

If we're looking for deeply impacting steppe influences, then it's with Yamnaya & possibly Cernavoda (although the latter hasn't yet been genetically characterised). As I've outlined previously, after 2500 BC, there's a retreat of steppe cultures in the Balkans. Yamnaya kurgans disappear, its Catacomb successor instead focuses to its southeast (as you & Dave point out here), & BB (wherever it came from) ends due to an expansion of Nagyerev from the south of the Danube (of course, at least as far as BB is concerned, much of it were assimilated, notable by the presence of L51 in Nagyerv itself).

Yes, you have groups like Noua-Cosgleni-Sabatinvka west of the Dnieper, which are thought to descend from Catacomb culture, but they're limited to Moldavia and parts of Transylvania; and they appear to me as Catacomb's ''poorer cousins'' rather than Myceneans on the eve of a conquest.
So im not sure why, but 2500-1900BC is a lull of steppe presence in Balkans; and only begins to change with the Srubnaya culture, which differ from Yamnaya -Catacomb (seem more ''settled'' than the proto-nomadic Catacomb), and their settlements reach the forest-steppe zone, and this is when that R1a-Z93 migrant appears in Bulgaria. So might relate to climate, at least in part.

On the other hand, after 2000 BC, it seems that it the predominant cultural influences moved northeast; notably the influence of Carpathian, eastern Halstatt, etc groups on the forest-steppe, the Bilohrudivka-Komaraov-Chernoles sequence in Ukraine.
But we have too few samples still for any hard hypothesis, still much to know

Folker said...

About datation the Anatolian branch's split (or any other branch), computational calculations are not reliable. Because any calculation is based on asumptions, like the one you made about Romance languages (which is a very bad choice to date a linguistic split, if I may. Latin was still an administrative language in many countries untill a recent date, and is still in somes, and many mutual influences are known between romance languages, to the point that they are still mutually understandable to some extent).
Problems are that those asumptions are difficult to validate. Another difficulty is about the "tree concept" itself. Interractions with languages of the same group or mutuel influence from another language family are usually underestimated.
Gray and Alkinson has been easily falsified (sufficient to say that they have given wrong results for some very well-known languages).
About Anatolian languages, Palaic and Hittite were heavily influenced by Hattic, but Hattic was itself influenced by Luwian (or a similar language).
To date, there is a large consensus to date the split between the late Vth millenium to the mid-IVth millenium. Roughly when the Yamnaya and related cutlures were emerging and when Steppe admxiture is found in the Balkans.

Nezih Seven said...

@Davidski can we consider Yamnaya_Bulgaria as a reliable steppe sample?

Lee said...


Well thank you for the statement that the models do give a reasonable fit.

I do not think I am thinking anything particularly odd.

Archeology seems to indicate that the Bell Beaker culture originated in Iberia-Independent of any Steppe genetic influence.

These Iberian beakers were moving by boats along both the Mediterranean, Atlantic and NorthSea. The genetic impact by them was not huge. They were trading with Corded ware peoples-whom subsequently appropriated cultural aspects of the Beaker people. IE-people have a long history of borrowing from other cultures.

I see no issue with the Iberians to have bought home wives from CW heavy area's and for having some families in various trade locations. Traveling salesmen/sailor are famous for having wives in every port.

I see no conflict with the Netherlands being a contact point brining in Western genetics into the Corded ware genetics of the region. The qpAdm certainly supports a mix of CW and Farmer/HG. My models support a local origin of the Farmer/HG ancestry but Iberian BB input is still a possibility---or both. I do think the models favor local HG/Farmer rather than external.

I have no problem with that corded ware population being from central europe/germany--being predominantly CW genetically. And subsequently taking over the bell beaker trade networks. That all makes complete sense to me.

My interest is where did the Iberian BB come from in the world? I think the data could support a Minoan connection.
I can model the Minoans as Anatolian-BA and WHG
fixed pat wt dof chisq tail prob
00 0 8 4.260 0.832939 0.956 0.044

As minoans were sea-going folk and they were attested all the way to Syria and perhaps SE Iberia--looking for a connection with the Bell Beakers to Minoans seemed logical to me.

They Iberian BB can be modeled as coming from the Minoans with different mixes of WHG-including ElMiron
fixed pat wt dof chisq tail prob
00 0 9 9.409 0.400417 0.182 0.818

but my best fit (out of many many population trials) was WHG and Beaker_Northern_Italy

fixed pat wt dof chisq tail prob
00 0 9 2.914 0.967607 0.150 0.850

So with this probability--it looked like the connection to Minoans was mediated via Beaker Northern Italy

So this begged the question of who were the Beaker Northern Italy related to via qpAdm
They sort of fit to Anatolia-BA + HG --HG as Villabruna came out as
fixed pat wt dof chisq tail prob
00 0 8 6.175 0.6276 0.186 0.814

Not a bad fit, but substituting in Minoan_Lasithi for Anatolia-BA provided a much better model

fixed pat wt dof chisq tail prob
00 0 8 3.945 0.862093 0.121 0.879

So the ancestry trail seemed to go--Anatolia-BA --> Minoan_Lasithi --> Beaker-Northern-Italy --> Iberia-BB with local mixing along the way.

That would all potentially bring some non-steppe associated CHG-like genetics across the mediterranean, during the bronze age.

Does not seem unreasonable genetically--though I do not know if the archeology would support the directionality of it.

Northern Italy could be a back migration from Iberia or something else...
Maybe Remedello culture mediated? They seemed to have non-steppe CHG as well...

This is all for fun--no dog in the fight either way

Gaska said...

@Lee Albee

No wonder you can model Iberia Bbs with northern Italy because the mitochondrial haplogroups of Via Guidorossi (Parma BB) are identical to the Spaniards although the Iberians are older. There are also coincidences with the few mitochondrial haplogroups that we know of Remedello, with which the Italian connection is more than evident. The archeology says exactly the same (Palmela spearheads and Ciempozuelos style in Liguria).

These connections are also evident with the deposits in Sardinia. Sicily, France and Morocco. You're right, the Iberian BBs were great navigators.

Regarding the Minoan connection, we have verified the uniparental markers existing both in Crete and in Iberia, the masculine ones do not coincide because they are J, and in the mitochondrial there are few coincidences (it is also true that the Minoans are more recent than the Iberians), We need more information about the Neolithic in Crete. And you know, the Minoans have some mitochondrial haplogroup identical to Maykop.

Thanks for paying some attention to Iberia.

Lee said...


I find the Iberian Area to be amazing

In the Paleolithic we have the Solutreans--the most advanced Paleolithic Culture in Western Europe. Some suggestions that the Solutreans may have been a combo of Iberian Paleolithic populations + some North African/near Eastern populations.

We have expansion out of the Iberian Glacial refuge in the Bølling-Allerød. Then a near complete replacement of that population during and after the younger Dryas by the WHG in most of europe. Though more of the genetics of this population were held over into the Neolithic.

I personally think that this turn over time helps explain the weird distribution of some mtDNA haplogroups across Europe.

We then have a Megalithic culture and then Bell Beaker culture coming from this area.

Just a repeated area of cultural innovations---a really cool place.

Makes me wish to be Iberian in genetics. I alas seem to not have much. Mostly a Steppe Invader :)


MomOfZoha said...

@Leron, Davidski, and others:

I don't want to unnecessarily extend this thread at this point, but wanted to say thanks for the reply/information.

Gaska said...

Lee, the old Iberia still has many things to say even in the matter of the invaders. You know that the Basques are more than 90% R1b and we speak a non-Indo-European language. We have no doubt about our origin.

JuanRivera said...

I'm ~72.5-78.5% Iberian, ~12.5 African-related and 10-15% Amerindian. Anyways, I should better focus on ancient DNA.

Davidski said...


Can we consider Yamnaya_Bulgaria as a reliable steppe sample?

Not really. That individual is basically 50/50 Balkans_ChL/Yamnaya.

JuanRivera said...

Here is what models show: the earliest Iberian beakers (and the earliest North Italian beakers) fit better as mixtures of Minoan_Lasithi/Anatolia_EBA, Iberia_Chalcholithic and French Neolithic. They have no steppe ancestry, as the G25 label no_steppe conveys. Later Iberian beakers, in contrast, fit best as mixtures of the earliest Iberian beakers and Dutch Beakers, who themselves are a mixture of Beaker_Central_Europe_no_steppe and CWC_Germany. Iberia_BA must have been even later, judging by slightly decreased steppe ancestry (from Beaker_Iberia's 25.83% to Iberia_BA's 20%, which is near modern central Spaniards' ~17%). Intriguingly, the MBA of Portugal has the exact same level of steppe ancestry as modern Portuguese (~15%).

Lee said...


Overall our models are pretty similar

Your model of iberians BB with S. Seem to indicate that we are looking at a minor influx back 2-3 generatiins.

Some states could be run to figure out the exact number of generations. But I am less familiar. With that usage and software.

Gaska said...

@Juan, Lee

The experts in handling tools using autosomal DNA are you. The important thing is to reconcile these models with what the uniparentals markers tell us, and also with archeology, linguistics and anthropology. Combining all this knowledge we will reach more convincing explanations.

Juan, I'm not sure which individuals you use to make those combinations and why you say some are earliest Iberian BB and others are not. You have to know very well the Spanish sites to know if they are BB sites or they are not, because that culture is very old in Iberia (there are deposits in Portugal and Galicia dated between 2,800-2,750 BC).What is certain is that with the data that we currently have the Iberian BBs have absolutely nothing to do with the CWC, except for that small autosomal "steppe" component, which on the other hand could enter Iberia by other ways (that is, not necessarily through the Pyrenees).Many Spanish and Portuguese geneticists say that this autosomal marker did not enter Iberia in chalcolithic (2,500-2,000 BC), but in the Bronze Age.

The maintenance of that small autosomal percentage in Spain and Portugal over more than 4 millennia is also very strange, because what we do know is that there were Central European migrations in the Iron Age (Urnfield, Celts), and probably also in the Bronze Age.

In any case, it is something that I wanted to discuss with you more carefully.

Un saludo

Aram said...


Thanks I see Your points. I agree that some time after Yamna there will be a resurgence of local EEF rich population. After all we have that sample from Iron age Bulgaria who is more Basal. This resurgence will obviously lead to changes in material culture, burial rites, etc.
But linguistics is a different issue. Language can come from one direction and material culture from a different and a new syncretic culture emerges. Which will have new qualities than his "parents".

That's why I wrote above that clarification of Greek's origin is important. My feeling is that I2c2 that was found in LBA/IA Armenia is a result of some kind EEF rich (with some steppe) pop expansion in Balkans or somewhere in Pontic region.

But overall I don't

Aram said...

But overall I don't believe this will result in a revolution in PIE origins question.


Having unusually high level of any Y dna in a population usually means extreme drift, which from linguistic point of view is not a good news. Chadic speakers also have very high level of R1b, but this doesn't mean their language came from Europe Highest level of R1a is in Kyrgyzstan. One of highest levels of G2 is found among Ossetians and the highest level of J2b gets an Austroasiatic speaking nation in India.. 75%.
Having high diversity of something is a much better option.

Gaska said...

@ Aram said..."Having high diversity of something is a much better option"

Better option, for what? To determine the origin of a language?

Basque is a language of Neolithic origin (at least) and is not related to Indo-European invasions. The simplest explanation is to look for the origin of R1b-L51/P312 in the Franco-Cantabrian region and its expansion to the rest of Western Europe through the BB culture. There are many geneticists, archaeologists and Spanish linguists who think that the theory of the Kurgans does not make any sense and that is what we want to prove. At the moment there is no L51/P312 neither in the steppes, nor in CWC, ergo it can only have western origin.

Aram said...


There is no much difference between L51 and the upstream L23 and M269. Where do You think M269 comes from?

Lee said...

Where does m269 come from? That is a good question.

Davidski said...

M269 is obviously from the steppe. That's where the oldest M269 is.

Gaska said...

@Aram, Lee, Davidski

The latest studies that have been done in Spain and France with more than 16,000 samples of P312 men indicate a possible origin of R1b-P297/M269 in Eastern Europe where he had to stay until the beginning of the Neolithic. However, I see different possibilities for the expansion of the lineage in Europe.

1- R1b-WHG originating in the Gravettian (Villabruna) expanding from Italy/Alps to the West (Ibousieres) and North-East (Latvian hunter gatherers)- Later migration of the Latvians more to the East (Ukraine, Russia) with development in the steppes of some branches of the lineage including L23-Z2013. It would be possible that R1b-L51 also originated in the steppes and then migrated to the West in very early periods (Khavlinsk, or even earlier, 5,500-4,500 BC). This model fails because so far there is no L51 in the steppes.

2- R1b-WHG originating in the Balkans/Rumania/Bulgaria- With movements in all directions (North-Latvia, East-Ukraine, West-Italy, Germany, France) in different historical periods and in different ways (land or sea). Some clades of mesolithic hunter gatherers migrated to the steppes (Samara) where they developed lineages totally different from the West. Probably R1b-V88 and M269 were incorporated into the dispersion of the Neolithic farmers from Anatolia, reaching Germany, France and Spain at the beginning of the Neolithic of those regions (5.500 BC)-Blatterhole, Kromsdorff, Quedlinburg, Els Trocs, El Portalón-It seems that none of those Neolithic R1b have steppe ancestry, then we would have to rule out their origin in the steppes.

I believe that both R1b-L51 and R1b-P312 are Western and no concrete origin can be ruled out for the moment (Germany, France, including the Franco-Cantabrian region).I do not see its relation with the steppes and less its relation with the expansion of the Indo-European language except in more modern historical times (Bronze Age, 2,000-1,100 BC).

Genetics gives us clues about the composition of different European Neolithic and Chalcolithic cultures and none of them seem compact enough to link a certain uniparental marker (male or female) to a certain language. I only see the exception of HapY-R1a whose relationship with the steppes and the CWC is more than obvious.If there is a male haplogroup that could disperse IE in Western Europe it was R1a, never R1b-L51.

The interaction of the BBC and the CWC (cultural, and to a lesser degree genetic) meant the dispersion of the steppe ancestry across the rest of the continent, although it is evident that it could also enter with some groups of Neolithic farmers (from Italy, Greece or even Anatolia).

Drago said...

@ Aram

“I agree that some time after Yamna there will be a resurgence of local EEF rich population.”

Well you cannot disagree, because the data is clear :)
What you & others (e.g. Reich) dimiss as ''bounce back of EEF'' actually occurs in a specific time frame & order, from c. 2500 during the Bronze Age, moving from Southeast Europe to northern Europe. It obviously means something, although what this is still unfolding.

“But overall I don't believe this will result in a revolution in PIE origins question. ”

I don’t think you there’s going to be any revolutions either, because there’s only a couple of tenable scenarios, but it’s details which matter about the final model (e.g. which areas around the Black sea the various waves of movement began)..

Drago said...

@ Juan

''Here is what models show: the earliest Iberian beakers (and the earliest North Italian beakers) fit better as mixtures of Minoan_Lasithi/Anatolia_EBA, Iberia_Chalcholithic and French Neolithic'''

Not at all.
Casting aside the fact that the samples labelled ''early Iberian BB'' aren't actually of Beaker provenance, they have no ''Minoan'' ancesstry or extra CHG ancestry. They're just local Chalcolithic groups, with little if any extraneous. I'm not sure where Lee is getting the exotic admixture from

Iberia_MN 92.3%
Trypillia:I1926 5.7%
WHG 1.7%
Samara_Eneolithic:I0122 0.2%
Aegean_N 0%
CHG 0%
d 1.76%

Lee said...


I replaced Beaker_North Italy with Remedello_BA in my model of BB_Iberia in qpAdm and got an even better fit

Remedello_BA + WHG
fixed pat wt dof chisq tail prob
00 0 8 2.968 0.936344 0.739 0.261

error is worse though, It looks like Remedello by itself gives a pretty fantastic fit with a probability of 0.89

So probably an error with the WHG or with my outgroups with this mix

So another Group high in CHG ancestry is showing up as a proxy population for Iberia_BB (excluding the "steppe" samples).

That of course begs what is in Remedello_BA?

So as with Northern Bell beakers I tested for Minoan_L and Villabruna as a possible source.

fixed pat wt dof chisq tail prob
00 0 8 3.492 0.89979 0.947 0.053

So a pretty reasonable fit--need to work on the error though (Std. errors: 0.085 0.085 )


Gaska said...


Of course it's interesting,

+ Emergence of the ideology of the warrior in the western Mediterranean during the second half of the fourth millennium BC- Christian Jeunesse (2.017). In the western Mediterranean, which covers, with various levels of intensity, the Italian península, Sardinia, Southern Alps and the south of the Iberian península and difusses towards the neighboring areas, notably the other regions of the Iberian peninsula and south-east of France, the ideology of the warriors emerges here far sooner than the Bell Beaker culture. According to the regions, it can be dated between 3.500-3.000 BC, and it marks the establishment of a new civilization which will bloom during the first half of the III Millenium. In the areas more strongly impacted by this movement, the appearance of the BB culture can be considered as a mere secondary event, which does not challenge the ideological and social bases. In Italy, the long lasting pre-BB period when the ideology of the warrior dominated, is known through the famous Remedello (3.350 BC), Spilamberto (3.350 BC), Gaudo (3.300 BC) and Rinaldone cultures. Grave goods in male graves from the dominating strata are composed of a dagger (in cooper or flint), a copper flat axe, a cooper quadrangular sectioned awl, and a more or less big amount of finely retouched flints arrowheads that probably accompanied a bow.

The same phenomenon has hit the Iberian peninsula, though to a lesser extent. The end of the IV Millenium is characterized by the spectacular development of metallurgy, the appearance of anthropomorphic steles and a whole range of copper objects that significantly resembles the one observed in Italy.

¿Origin in the Steppes?- The ideology of the warrior, often still considered to have been created by the Late Neolithic II “Mega-cultures” (Corded Ware, BB culture),is definitely present in Italy in a completed form as early of the third fourth of the IV Millenium- The development or the renewal of the cooper metallurgy, the innovative intentional use of the copper –arsenic alloy, the single male burial containing warrior grave goods, the higher social and symbolic value of the weapons and the production of anthropomorphic steles. From my point of view, the geographical scale and the homogeneity of this phenomenon are decisive arguments against the sometimes put forward hypothesis of local evolution from the various cultures that were present in the concerned region before 3.500-3.000 BC. Harrison and Heyd describe a migration from East to West, prior indeed to the BB diffusion, but according to them notgoing back beyond 2.900 BC. Yet, the weak point lies in totally blanking the fact that the first occurrences of the “steppe set” in Western Europe are a great deal prior to the Yamnaya event (3.300-2.500 BC), since they are present as early as the last third of the IV MIllenium in the areas mentioned above. We are thus convinced of the existence of a first wave of steppe influences during the last centuries of the IV Millenium (Maikop Culture, Anthropomorphic steles, 3.800-2.500 BC), which constitutes the opening event of an expansion that afterwards goes on all along the Late Neolithic I horizon (3.500-2.500 BC).

Aram said...


R1b/WHG sounds interesting but don't stick well with some aDNA data. Do You know that of the oldest case of R1b PH155 (which is the first branch of R1b) is found in Xiongnu.
The oldest case of V88 is from Iron Gates Mesolithic. Those samples had extra ANE compared to other WHG follks.

Aram said...

Also how R1b can have different origin from R1a? They had a common ancestor after LGM. So they started from the same autosomal context.
And current evidence point to the EHG origin of R1.
Post LGM colonisation of East Europe and it's neighborhood.

Aram said...

Also V1636, M73, and M269 all had a common ancestor at 15.000 years ago. And now try to find were the oldest aDNA of this SNPs are found. I mean undisputed aDNA.

As strange and unbelievable it can sound it is in East Europe.

Gaska said...


I think you're mixing different debates and I'm not sure what you mean.

1-It is obvious that R1b is a lineage of WHG (at least from the Gravettian), many geneticists believe that it can not be older in Europe than the one discovered for Villabruna (Epi-gravetian), but this is a question that depends on new discoveries in both Western and Eastern Europe. Villabruna- Kit Number-M236020- Populations- ANE- 7,16%, ASE- 0,52%, WHG-UHG- 89,88%, East-Eurasian - 0,97%, West African- 0,36%, East-African- 1,12%,ENF-x. It is assumed that if R * originated in Siberia all of its close descendants have to carry the autosomal signal ANE, so it is therefore normal for Villabruna to have it.

2- I'm not sure which is the oldest case of R1bV88 but obviously the Balkans is a possibility. That agrees with the hypothesis of its migration towards the West during the Neolithic (Els Trocs, Germany etc), and towards the East-
SC1- Romanian Hunter Gatherer-HapY-R1b1a/2-V88 or Pre-V88-6.814 BC
I4666- Serbia, Lepenski-Vir (6.067 BC)- HapY-R1b1a/2-V88-Mit-H40.
I1734- Ukraine, Vasilievka, (7.252 BC)- Hap Y- R1b1a/2-V88).

I think there is consensus on these samples, I do not know if you will know any more old ones.

3-Obviously the origin of R1a and R1b is the same, but its history is very different.R1a expanded and could even originate the oldest cultures of the steppes, and from there it expanded to central Europe and India. I think that anyone who understands something of genetic will be in agreement with this. I believe that it is a lineage linked to the Indo-European. R1b is totally different, and the mistake of thinking that the origin of a haplogroup means that its brothers have to have the same origin is very common between professional and amateurs geneticists .That leads people to think that since Z2103 is common in Yamnaya, L51 also has to be there- grossly mistake- It is as incongruous as thinking that L21 has its origin in Iberia because Df27 is mostly Iberian.
So they started from the same autosomal context, but this does not mean anything because the autosomal signal can be lost in a few generations and you have to keep in mind that in this case there are thousands of years away.

4- "And current evidence point to the EHG origin of R1"-

"A population of mixed WHG and Upper Paleolithic Siberian ancestry (related to the Mal’ta and Afontova Gora specimens from Lake Baikal ~24-17kya) are attested in European Russia ~8kya"

Then EHG is partially WHG, and with that mixture may have arisen R1b in the east.
in fact I have already said many times that the mitochondrial haplogroups of the Latvian hunter-gatherers (80%WHG/20%EHG) are very similar to those of the Ukrainian EHG, although the Latvians are always older.

4- I am not an expert in PH155, and I do not know the antiquity of that haplogroup in China. I think that this SNP has an antiquity of 7-8000 years, I would have to consult it.

This conversation has started because according to the latest work I have read from prestigious Spanish geneticists, they seem to agree on an origin of R1b-M269 in Eastern Europe (Balkans, Bulgaria, even the steppes), but it is also obvious that it is a Neolithic migration to the West, and therefore the possibility of an origin of L51 in Central Europe or even Western Europe is certain.

There is data that only a blind man can deny- Yamanya is R1b-V1636 and R1b-Z2013. Where is L51?

R1a is in the steppes for thousands of years and its connection to the CWC is evident, you can not mix the history of both haplogroups even if they are siblings, because R1a has a much more consistent historical behavior than R1b.If you mix them, you not only distort the story of R1b, but you complicate that of R1a.

Drago said...

@ Aram

“Do You know that of the oldest case of R1b PH155 ”

Did you mean the most divergent ?
Funnily enough it’s in a man of Italian background

Gaska said...

@ aram

It is important to take into account the supposed periods of formation of the different SNPs in the R1b lineage

R1b1a/1a-P297- Formed (13.600 BC)- TMRCA (11.300 BC)- We have seen that the Villabruna cluster (and associated Mesolithic R1b-M343 lineages found from west to east Europe) might be the representatives of the expansion of West Hunter-Gatherer ancestry, displacing or admixing with the existing population of Western Europe- R1b-M269 lineages may have thus brought WHG ancestry to the north Pontic steppe during the Mesolithic

R1b1a/1a2-M269- Formed (11.300 BC)- TMRCA (4.400 BC)-How do you interpret those 7000 years of difference between formed and TMRCA ?

It seems clear right ?. R1b M269 was paralyzed for thousands of years, and then there was an explosion of their L51 descendants in western Europe.

Gioiello said...

@ Diego

“R1b M269 was paralyzed for thousands of years”.
It means only that the expansion was recent (around Italy and the Balkans) and doesn’t say anything about its origin, its bottleneck, thus other haplotypes may give the needed information about the place of origin, and they are all in Western and Central Europe (of course I think that the place of origin was around the Southern part of the Alps) and we have infinite hints about that and I am speaking of them from more than ten years. I wrote perhaps 15000 letters about that. We need proofs and I think that they are in Italian aDNA next to come. Of course Iberia had R1b only from the migration from Italy beginning from 7500 years ago, the migration Zilhao spoke about firstly.

Aram said...

I was meaning oldest ancient DNA. PH155+.
The oldest PH155 in Europe is found from a Gepid who was a mixture of Hun and locals. Other cases of PH155 was found in Tian Shan Huns.

What a modern Italian do not change much. Because You need to have it in ancient Italy, older than those Huns.

And I don't believe that R1 is from Siberia.

Origins of PH155 is very very important.

Gioiello said...

@ Aram
“I was meaning oldest ancient DNA. PH155+.
The oldest PH155 in Europe is found from a Gepid who was a mixture of Hun and locals. Other cases of PH155 was found in Tian Shan Huns.

What a modern Italian do not change much. Because You need to have it in ancient Italy, older than those Huns.

And I don't believe that R1 is from Siberia.

Origins of PH155 is very very important”.”

I don’t think that R-PH155, and also R-PH155-M335, are surely from Europe. These are subclades too recent and may have come in Middle Ages with people from East as other hg R surely come with Turks or others (they are the descendants of some R-Z2103 subclade of Yamnaya and migrated to Central Asia and entered the Turkish pool, but that doesn’t mean that R-L23 was born in Asia and neither in Yamnaya). Certainly Western European R-M73*, separated from Easterrn European and Asian R-M73-M478 about 9000 years ago, didn’t come to Western Europe from these Eastern haplotypes, even though I said that we cannot exclude an old origin from the R1b from Baltic (but they were anyway the Western European hunter-gatherers descendants after the Younger Dryas).
About the other haplotypes go and read my thousands of letters, also on this blog before my banishment.

Bobo said...

It’s just as easy to explain the proto-Armenians as having come from the Armenian Highlands/South Caucasus as it is from the West considering that there was an influx of migration from NE Turkey/modern Armenia as far as Elazig during the Bronze Age Collapse (we know this from the prevalence of the so-called Transcaucasian pottery). This entire region corresponds with historic Armenia. The three groups that were mentioned in relation to the spread of “Transcaucasian” pottery culture were the Urumu (possibly Armenians), Mushki (possibly Armenians, or a related people), and the Kaska (Apishlu).

Bobo said...

Catacomb>Early Kurgans>Trialeti-Vanadzor (proto-Armenian)>Hayasa (Urumu (Armenians), Mushki (Armenians or Armenic), Kaska (Caucasian))

Bobo said...

They have found identical grave artifacts in Trialeti-Vanadzor as in Mycenaean graves. This means they were either in contact with one another or they were related culturally (these options are not mystically exclusive, however). We know that the Myceanaeans have some ancestry from the Armenia area. This is one of the reasons I think Greeks entered the Balkans from Asia Minor (maybe through Troy or across the Aegean, or maybe both) as opposed to entering from the north of Greece. Essentially what I’m arguing is Eric Hamp’s model of migration—Armenians and Greeks separating in the Caucasus, Armenians staying and Greeks moving west.

Bobo said...

But why focus on the Kaskas when the Urumu (Arama?) were mentioned around 1200 BCE as well, in conjunction with the Eastern Mushki, who are believed to have been Indo-European?