search this blog

Thursday, December 13, 2018

Italians are interesting people (Raveane et al. 2018 preprint)

Over at bioRxiv at this LINK. As far as I can see from skimming through the preprint, it's a thorough effort with conclusions that make good sense. But, needless to say, it'll be very useful to plug the dataset from this study into the Global25 to see what these new Italian samples are really made of. Here's the abstract, emphasis is mine.

European populations display low genetic diversity as the result of long term blending of the small number of ancient founding ancestries. However it is still unclear how the combination of ancient ancestries related to early European foragers, Neolithic farmers and Bronze Age nomadic pastoralists can fully explain genetic variation across Europe. Populations in natural crossroads like the Italian peninsula are expected to recapitulate the overall continental diversity, but to date have been systematically understudied. Here we characterised the ancestry profiles of modern-day Italian populations using a genome-wide dataset representative of modern and ancient samples from across Italy, Europe and the rest of the world. Italian genomes captured several ancient signatures, including a non-steppe related substantial ancestry contribution ultimately from the Caucasus. Differences in ancestry composition as the result of migration and admixture generated in Italy the largest degree of population structure detected so far in the continent and shaped the amount of Neanderthal DNA present in modern-day populations.

Raveane et al., Population structure of modern-day Italians reveals patterns of ancient and archaic ancestries in Southern Europe, Posted December 13, 2018, bioRxiv, doi:

See also...

Greeks in a Longobard cemetery

Migration of the Bell Beakers—but not from Iberia (Olalde et al. 2018)

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


Cpk said...


Davidski said...

Yeah, maybe Etruscans in part, but in any case, this CHG-related signal can probably be largely explained by the so called Iran-related gene flow into the Mediterranean from the east during the Bronze Age that we discussed here...

How should we interpret the movements of people throughout Bronze Age Europe?

Bob Floy said...

" Iran Neolithic (IN) ancestry was detected in Europe only in Southern Italy."

Out of respect for the blog I won't use any emojis, but it sure was tempting.

Gökhan said...

i like the title. "italians are interesting people". lol.

Ric Hern said...

Yes that Iran Neolithic jumped at me like a Tiger. That one sentence made my day...

Bob Floy said...


I'm half southern Italian myself, so it's extra hilarious for me in the middle of this debate to find out that I personally have Iran neolithic ancestry, while we know that most European descended folks(like the Max Plank researchers, probably) do not. Made my whole week. Also this paper is a great read just in general, and long overdue I think.

Matt said...

Nice new samples, I like seeing these Chromopainter driven papers as well as a change.

That said, not sure their CP/NNLS (ChromoPainter non-negative least square algorithm) fits really make that much sense. Considering the distal fits, you only have about 25% EHG in NW Europe, but no CHG... Where conventional models go with NW Europe as about half Yamnaya, this suggests quarter EHG with none of CHG that contributes to Yamnaya.

While EEurope has implausible amounts of EHG (40%) and very different EHG:WHG ratio. EHG only about 5-10% in Spanish (WEurope) clusters, peaks in Basque cluster WEurope2, which lacks CHG.

Even the models with proximal clusters that allow Steppe_EMBA ancestry, you get 30% Steppe_EMBA in NW Europe with only the E European cluster (e.g. EEurope2, largely Polish, for'ex, getting close to 50%).

(Figs for levels:

Iran_Neolithic in cluster analyses but no Levant_Neolithic is an odd choice.

Sardinian cluster differences are cool to see; the cosmopolitan / outbred Sardinian cluster (look at self-copy levels) is a little closer to everyone else, but still pretty similar to the inbred cluster.

Neanderthal stuff not so interesting. Should use direct ratio, not indirect.

ǵenh said...

Etruscans have nothing to do with this CHG, CHG is stronger in the areas of Italy where the Etruscans have never been there. Perhaps it is appropriate to start talking about prehistoric cultures.

Matt said...

Genotype PCA and linked chunkcount PCA from paper, with dendrograms for clusters and their composition:

Mostly similar but differences:

1) As normally the case in these, Sardinians and Basques (cluster W Europe 2 here) are much closer to Western Europe in the first two dimensions of chunkcount PCA than genotype.

The correlation between the genotype and chunkcounts PCA breaks down more for them, for whatever reason (perhaps their haplotypes are very different and can't be captured by PC1 and PC2 of chunkcounts, or perhaps they're not very different, even though the genotypes are).

2) EEurope 5, mostly composed of Hungarians, Slovaks and Slovenes, is obviously intermediate in the chunkcount PCA between NWEurope 5 - German and Scandinavian - and Balkan clusters - Croat, Serb, Montenegro, Kosovo, Bosniak. Alternatively being intermediate NItaly6 - NNE Italy - and EEurope2 - Polish(Estonian?), Russian, Ukrainian - also works.

3) WAsia7 - largely Turkish - markedly closer to Balkans and SEEurope in genotype than chunkcount. Low recent geneflow between Anatolia and Balkans/SEEurope (though WAsia8, shows there must be some, as does the stream away from North Caucasus Caucasus7 and Caucasus2 to EEurope) but links in older components?

4) Sicily7 cluster seems closer to Sardinian and Spanish cluster in chunkcount compared to genotype.

(Parenthetically, considering again the NE Italian cluster, still seems like a shame that the highly isolated samples from have still never been analysed in light of ancient dna and formal stats that minimize the role of drift).
Btw, does anyone else find the supplementary material tables to be missing from biorxiv? E.g. going back to Mathieson's paper from 2017 - Annoying if so.

Ric Hern said...

@ JuanRivera

The reconstruction of horse is interesting since there are so many ways. Some IE languages seem to use words that referred to Stalliion, Mare or Foal etc. and applied there words to the collective word for horse. So IE could have had many different words which evolved differently in different languages and makes it more difficult to see the connection to PIE...

Simon_W said...


The obvious problem of this Theory is: Anatolia_BA-related ancestry peaks in the South Italian (dark brown) and Sicilian (yellow) main clusters. And it's higher even in the south-central Italian (purple) cluster (which peaks in Umbria and is strong in Latium and the Marche) than in the Tuscan clusters. However, the Etruscans lived most of all in Tuscany and northern Latium (while the Etruscans of the Emilia were assimilated by invading Gauls). The overall distribution pattern of Anatolia_BA-related ancestry in Italy has been known for years, but this paper complements and confirms it with further data points. So in any case, it strongly looks like the Etruscans had less Anatolia_BA related ancestry than Italic speaking populations like the Umbri and the Latins, and the South Italian Oscans and the Sicels etc. had the most. Granted, the Tuscan clusters still have substantial Anatolia_BA-related admixture, that's true, in theory this could be Etruscan, but it could just as well be a diffusion from the south, it definitely isn't an Etruscan-centered phenomenon.

Leron said...

JuanRivera: I don't see a political reason to maximize CHG in relation to Steppe ancestry in these studies. Maybe only in the old days or for those who still believe in fringe theories like "posited" Armenian homeland for IE. Everything is clearly signaling Caucasian speaking populations that moved around the Mediterranean. Perhaps attested in Greek sources as the various Pelasgian groups, and archaeologically related to the pulse that brought forth Khirbet Kerak and Kura-Araxes type remains in Cyprus during the early Bronze Age.

Leron said...

ǵenh: Keep in mind that massive upheavals brought an end to the Etruscans, in addition to populations that kept coming from the north almost without stop. Nowadays, can we detect Hittite ancestry in Turkey or Tartessian ancestry in Spain? They may not have left very few direct descendants and most of the related people who survived have been displaced. What source for Etruscans would be left then? Steppe is out of the question, so that leaves few remaining choices.

Gabriel said...


1. To be more inclusive/less Eurocentric (maybe also avoid being seen as racist, or even taunt Euros)

2. Association with the roots of agriculture and ancient civilizations, which seems more magnificent and richer than being descended from seemingly primitive pastoralist nomads from the Russian steppes (sometimes akin to hyperdiffusionism

3. Inability to abandon Hajji Firuz, the idea of Maykop as IE or the Anatolian and Armenian hypotheses, in spite of the genetic evidence

Simon_W said...

The Italian structure as seen in the clustering corresponds nicely with Matt's PCA:
Umbria, Lazio and Marche are very close to each other. Tuscany is different, more northern. The Emilia is closest to Liguria. Bingo, I'd say!

But a susprising, to me exciting new finding of the preprint is the complete absence of Anatolia_BA related ancestry in the Northeast Italian cluster (beige). Only the Aosta Valley cluster (NItaly1) and the Northeast Italian cluster (NItaly6) are without Anatolia_BA-related admixture.

Another thing: Matt's PCA shows the Romagna quite far from the Emilia and very close to Umbria, Latium and the Marche. So judging from this, the south-central Italian (purple) cluster should also be strong or predominant in the Romagna, and the authors of the preprint sampled the Emilian-Romagnol sample in the Emilia.
However, as far as I can tell from my own DNA, there seems to be variation in the Romagna, whether on a small-scale geographical or individual level, I don't know. 1/4 of my ancestry, or 25.7% of my genome is from Cesena, Meldola, Montiano and Forlimpopoli (the remainder being German and German Swiss). And this is what I get with the Global25/nMonte approach:

Considering that even the Emilian/Ligurian cluster has some Anatolia_BA, it seems to me that the Northeast Italian cluster extends somewhat southwards into the Romagna, at least in individuals. But there definitely are many Romagnols with more Anatolia_BA (including one DNA relative of mine).

Synome said...

At the risk of derailing into the Etruscan debate:

I used to support a post Neolithic origin of the Etruscans in Italy (the Anatolian migrant theory).

However, a lack of genetic evidence for this hypothesis (no distinctive post Neolithic Anatolian signal in Tuscany or north Italy), along with the issue of parsimony in linguistics (most languages shown to related to Etruscan are nearby, and Lemnian is not definitely established as native to that island) has lead me to now support the indigenous theory. Of course the ancestors of Etruscans would have indeed originally migrated from Anatolia...but as EEFs, not Bronze or Iron age migrants.

Etruscan may just be a European Neolithic farmer descended language, like Basque and pre IE Sardinian. If Basques can keep their language while having their Yhg lines replaced and autosomally coming to resemble their IE speaking neighbors, what's preventing a similar process happening with Etruscan?

Grey said...

i wonder if there was a phase just after mining was figured out where populations who knew how to do it spread widely as a minority by hopping from one soft metal site to the next with Iberia being the final destination - lots of soft metal.

wild speculation
- Maykop founded by outsider gold prospectors becoming the source (or a source) of later CHG miners?

more wild speculation
- two sets of soft metal mining associated populations from the Caucasus region reaching Iberia, a CHG like population taking the southern maritime route (Cyclops legends, Nuragic culture, sea peoples?) and a more EHG like population taking the Volga route and then down the Atlantic coast (Thor's (blacksmith) hammer, nordic "dwarves")?


i think the Basques unusual mixture is possibly a clue there may have been two IE waves - an earlier more artisanal one somehow connected to trade (miners, horse traders?) and a later tribal migration one.


i think the plague idea if correct could neatly explain the HG resurgence, sequence

1) farmers expand to the edge of viability pushing HGs to the fringes
2) farmers catalyze border HGs into herders
3) ex-HG herder's animals create a plague
4) ex-HG herders move back onto their old territory


another variation on the idea of miner/trader/artisanal type migration that might fit southern Italy could be obsidian maybe - i assume that would be much earlier though.

Simon_W said...

Keep in mind that massive upheavals brought an end to the Etruscans, in addition to populations that kept coming from the north almost without stop.

Ha, are you suggesting there is substantial Longobard or Ostrogothic admixture in Tuscany? Or Gaulish even? There is no evidence for such a thing! Furthermore the Tuscan/northwest central Italian part of Italy is clearly distinct from the other parts of central Italy, both in genome-wide comparisons and in the yDNA. What this suggests: The Etruscans haven't been replaced by Romans. Otherwise there wouldn't be that clear difference we see today.

Where the Etruscans might have come from? You probably know there were related languages in the Alps with Raetic, and in insular Greece with Lemnian. So either they came from around the Alps, maybe with the pile dwellers in Northern Italy, or via a maritime movement from Greece. But it would have to be from a part of Greece where the Mycenaean culture was virtually absent, because there's hardly anything Mycenaean (pottery I mean) in the parts of Italy that became Etruscan speaking.

Simon_W said...

^that was @Leron BTW.

Simon_W said...


"most languages shown to related to Etruscan are nearby, and Lemnian is not definitely established as native to that Island"

What do you mean by "not established as native to Lemnos"? There are more than 100 (albeit mostly short) Lemnian inscriptions from various places in Lemnos. And the non-Greek inhabitants of that island have also been mentioned by Greek historiographers like Thucydides. So at least they had a firm presence there, this wasn't just a shipwrecked expedition. When exactly they first appeared there is another question. But if that makes them potentially non-native, then the same applies to Raetic and Etruscan.

And besides, there are not many more languages definitely shown to be related to Etruscan. Actually I don't know of any besides Raetic and Lemnian. Some fragments from around Pesaro that had previously been classified as North Picene were shown to be Etruscan-related. But they are so short that you can't establish a whole distinct language with them.

Ric Hern said...

@ JuanRivera

The Crabapple is basically indigenous to Acid Soils like the Lower Don all the way to the Balkans....

Synome said...


Sorry if I wasn't more clear about Lemnians. "Native" is a loaded word. I meant to imply the possibility that they may have arrived in Lemnos during the Bronze age or early Iron age, rather than departing from the region.

I've become more attracted to the idea that the Lemnians were descendants of mercenaries or Sea Peoples who settled on the island.

If you know of evidence against that scenario I would definitely be interested in seeing it.

Ric Hern said...

@ JuanRivera

The Apple which got stuck in Snow Whites throat was a Crabapple I think. Very Dry...Heheheeh.

Ric Hern said...

@ JuanRivera

Particularly Malus sylvestris.

M said...


The inscriptions of Lemnos do not even reach 10, let alone if they are more than 100. They are two main inscriptions and 5 or 6 very short inscriptions. The alphabet used in Lemnos for these inscriptions is the Western Red Type, which has been used mainly in Italy and Greece and this is clearly in favor of an arrival from the west. While the Rhaetian inscriptions are now about 300, the Etruscan ones are 13,000.

Grey said...

"Thor's (blacksmith) hammer"

i've been thinking smithing hammer for a while as the handle is inconveniently short for a weapon (recognized at the time hence a specific legend to explain why it's too short) plus the connection with lightning bolts maybe coming from sparks but on reflection lump/club/thor hammers are a lot heavier than smithing hammers (video on smithing hammers)

and are generally used to break up bricks, stone etc

so maybe not smithing but stonework - hammering wedges to crack rocks?

(might still get sparks?)

just a thought


another interesting titbit i didn't know till just now which may or may not be relevant

"Club hammers are common on the British inland waterways for driving mooring pins into the towpath or canal bank."

Toby_P said...

Sorry who or what is Molgen ?

Davidski said...

Alright, back on topic please. Italy...

Gabriel said...

In that map it looks like Northern Italians have more steppe than Iberians, the same as Irish and the Irish the same as or less than England and South Slavic speakers. Is that the case or no?

Ric Hern said...

Interesting that Greece looks more Steppe like than Italy.

Drago said...

In light of prevailing data spots I find it hard to believe any of the languages of Europe are Neolithic. I think they're all Expansions which are from after the Copper Age, and then further expansions through the Metal Ages. Despite early Farmer ancestry still being a dominant genetic component, there was too much local shift to posit any continuity with the Neolithic. When societies collapse or are replaced, there is little incentive for newcomers to adapt by-gone languages. Instead of static views of continuity, I think the reality was one of continued replacement and adstratum impacts.
My personal view is that northwestern and western Europe were Indo-Europeanised fairly recently. I think bloggers like Andrew (D.F.T.I) have made neat summaries about this.

Drago said...

I very much look forward to data from Italy. I find it surprising the absolute lack of DNA from it (although I suspect this is mostly due to local beuracracies than anything on the part of geneticsts).

Ric Hern said...

@ Dragos

"My personal view is that northwestern and western Europe were Indo-Europeanised fairly recently"

From where and when do you propose this happened ?

Ric Hern said...

@ Dragos

If you mean by Steppe related Central European Bell Beaker People around 2000 BCE, then yes.

Drago said...

Ric- I think this occurred as waves from central Europe with cultures like Urnfield, although details will have to wait until we get more data from the Metal Ages.

I just happen to have a difficult time accepting that Iberian languages are the remains of Neolithic/ pre-Beaker cultures given the extent and character of population replacement we are seeing in western Europe. Yes, this includes Iberia, too.

Beaker_Southern France 56.7
Iberia_CA 34.4
(distance 2.69)

Compared to the commonly quoted figure
Iberia_CA 59
Beaker_CentralEurope 41
(distance 2.81)

You have a hefty population replacement, elite conquest and cultural change all rolled into one.

Slumbery said...

"While EEurope has implausible amounts of EHG (40%) and very different EHG:WHG ratio."

I used the online G25 nMonte runner to test multiple European populations for CHG + EHG + WHG + Barcin_N. I am telling this as an answer to your comment, because I got totally unrealistic results and the reason behind it might be connected to unrealistic results by others.
I got unrealistically high EHG to CHG ratios all around Europe with this test. For example:

Barcin N: 49.17
CHG: 9.17
EHG: 35
WHG: 6.67

This would suggest that Hungarians have a lot of EHG ancestry other than what comes from steppe ancestry. WHG is also rather low.
I tried to change what represents WHG and even tried with LBK N and even Levant N instead of Barcin, but nothing solved this.
My conclusion for now that the CHG sample is too old. EHG itself have CHG ancestry that is younger and closer to the later steppe CHG that this CHG sample. Because of this the algorithm attributes a considerable part of CHG to EHG. Lowered WHG is a side effect of this, some of it also goes to EHG for balance.

And a side note for the CHG source in the steppe debate: I tried to use Ganj Dareh N instead of CHG, because it is younger, but the results were even worse.

Gabriel said...


Where do you think R1b fits in this? It’s impossible that it is Basque unless Vasconic languages were spoken in the steppe where R1b-M269 was found.

R1a is not the sole Indo-European marker and wasn’t the source of IE in Urnfield.

I thought we were over this mentality...

Drago said...

@ them meee
No I don't have a unilateral Hg mentailty, that's exactly my point.
For example, I think most people would look to R1b-U152 and R1b-U106 having a special part in Celtic & Germanic, resp (although not exclusively), although details are still waiting...

Gabriel said...


I would also consider L21 to be hefty in some Celtic groups, though widespread and disparate (found in British and Irish Beakers and in Scandinavia, so in the latter case probably in Germanics too)

I find it impossible for Basque to have spread with R1b-M269 unless they were absorbed by and then later brought by steppe-derived groups, or some linguistic shift towards Basque happened. No way that Beaker migration was associated with Basque.

Plus the whole R1b/R1a thing is cause back in the day some people thought Urnfield Indo-Europeanized Europe via R1a groups causing a language shift among Western Europeans rich in R1b which was thought to be native to the region or even associated with Basque, because modern distribution. Some (mostly Slavocentric) users in here still believe this. So to me it’s more than just haplogroups.

Ric Hern said...

Wonder what happened there at: Ancient ancestries in Western Eurasian modern-day clusters and Italian ancient samples. Northern Italy 4 when you compare figure A and D ?

Ric Hern said...

@ Them meee

Yes, Basque only make up +-1% of the total R1bs in Western and Northwestern Europe....

Davidski said...

@Them meee

Where do you think R1b fits in this? It’s impossible that it is Basque unless Vasconic languages were spoken in the steppe where R1b-M269 was found.

Not to thrown any fuel on the fire, if there is one, but you might find this interesting...

Advances in Proto-Basque Reconstruction with Evidence for the Proto-Indo-European-Euskarian Hypothesis

Simon_W said...

I think maybe the biggest flaw of this preprint is that they didn't include Natufians and the Bronze Age Levant in their CP/NNLS models as ultimate and proximate sources, respectively. I guess the rather high North African admixture in the Sicilian clusters may in part be caused by this. Judging from their Globetrotter Analysis, the North African admixture in Sicily goes back to the time of the Arabic invasion. But the same analysis also shows the Levant/Middle East as a proxy for the 2nd source, that is, the locals, although South Italians are the better proxy for that source. Anyway, Sarno et al. 2017 have shown quite clearly that there is substantial Near Eastern-like admixture in Sicilians. Let's hope Raveane at al. fix this in the definite version.

Also, I'd like to know what the fits are for each model. Because in Fig. S9 they use alternative sources for the extra-CHG input, which yield somewhat different models. For instance, using Minoans as a source instead of Anatolia_BA, two of the Tuscan clusters get rather high admixture from that, on a par with South Italians! Also the Liguro-Emilian cluster NItaly3 gets much more substantial admixture from that source. So that may be an Etruscan signal, if the latter had anything to do with pre-Greek Greece. But for that reason it would be interesting to compare the fits of the models. I guess they may be in the supplemental tables, but can't find these.

Simon_W said...


"I've become more attracted to the idea that the Lemnians were descendants of mercenaries or Sea Peoples who settled on the island.
If you know of evidence against that scenario I would definitely be interested in seeing it."

No, I don't, and I also consider this as a possibility for sure. Especially considering the archaeological evidence for Italian mercenaries in LBA Greece.
For me the question about Lemnian origins is just not as settled as it apparently is for some Italian users. Wishful thinking may be at play.


"The inscriptions of Lemnos do not even reach 10, let alone if they are more than 100. They are two main inscriptions and 5 or 6 very short inscriptions."

You're quite wrong. You forget the dozens of graffiti and dipinti from Efestia and Chloi, see:

Alessandro della Seta, Iscrizioni tirreniche di Lemno. In: Scritti in onore di Bartolomeo Nogara, raccolti in occasione del suo LXX anno. Città del Vaticano, p. 119–146 and tavv. XV-XVI.
Luigi Beschi, Atitaś, in: La Parola del Passato 51, 1996, p. 132–136.
Luigi Beschi, Nuove iscrizioni da Efestia, Annuario della Scuola Archeologica di Atene 70–71 (1992–1993) [1998], p. 259–274.
Luigi Beschi, Il Cabirio di Lemno: testimonianze letterarie ed epigrafiche, in: Annuario della Scuola Archeologica di Atene e delle Missioni Italiane in Oriente, p. 74–75, 1996–1997 [2000], p. 7–145.

"The alphabet used in Lemnos for these inscriptions is the Western Red Type, which has been used mainly in Italy and Greece and this is clearly in favor of an arrival from the west."

Ha, so you think their usage of an alphabet that was also used in Greece suggests that they didn't come from Greece?

Simon_W said...


"I think bloggers like Andrew (D.F.T.I) have made neat summaries about this."

I'm not familiar with his latest musings, but he has put forward some strange ideas in the comment section here. Like that the Indo-Europeans, which he equated with the Corded people, were cremating, while the non-IE Bell Beaker people were not. Indeed Andrew claimed that the spread of cremation we see in LBA Europe was a sign for the expansion of Corded Ware-derived IEs. But as I said before:
The Corded people inhumated their dead, they did as a rule not practice cremation! Their typical burial rite in central Europe included the orientation of the dead body along an east-west axis, the females on their left side with the head towards the east, the males on their right side, with the head towards the west, both sexes looking towards the south. In lesser Poland and the Ukraine on the other hand, the dead Corded people were oriented along a North-south axis, but also with the females on the left side and the males on the right side, with both sexed facing towards the east. And that's BTW rather similar to the standard burial rite in the steppe admixed eastern Bell Beakers, they also oriented their dead along a North-south axis, the difference being that they had the males on the left side and the females on the right.
I think Andrew was misled by the name "Urnfield culture". Which is a historically established designation for a certain complex of cultures, of which cremation on its own isn't its distinctive feature. Because cultures who predominantly cremated their dead already existed long before the Urnfield and Lusatian cultures. In the Carpathian Basin, cremation was already widespread and common in many cultures during the EBA! And as a matter of fact, there was no Corded Ware in the Carpathian Basin, only some Bell Beaker influence. Cremation was predominant in the following EBA cultures of the Carpathian Basin: Nagyrev, incrusted potters aka North Pannonian culture, Nyirseg-Zatin, Hatvan, Vatya (the inhumation graves where we got DNA from are mostly from late Vatya) and Vattina. Cremation and inhumation were equally common in Verbicioara and Kisapostag. These are all EBA cultures who date long before Lusatian culture and Urnfield. So the origin of the LBA central European cremation fashion has to be sought in the Carpathian Basin and not among the Corded people.

Simon_W said...

What regards my own above musing that the northeast Italian cluster NItaly6 extends into the Romagna judging from my own DNA - I've changed my mind about it. It's probably rather the case that the Liguro-Emilian cluster NItaly3 extends into the Romagna. The commonalities in dialect and the distribution of monophyletic surnames clearly favour this view. Now cluster NItaly3 does have some Anatolia_BA related admixture, but on a low level. And I guess the Yamnaya_Samara in my Global25/nMonte model may have a bit more CHG than the Steppe EMBA groups that are part of my ancestry. In other words, the Yamnaya_Samara probably eats some of my Anatolia_BA related admixture, leaving only a bit of Natufian not devoured.

Simon_W said...

Well, a lot has been said about Etruscans, Basques, and Europeans in general, but this preprint also offers some interesting insights about Italian DNA that haven't been commented yet:

Figure 1E for example, the EEMS analysis of southern Europe, shows some interesting "highways" of migration from the east to Italy. Mainland Italy is connected with mainland Greece, whereas Sicily is connected with Crete, Cyprus and the East Med. Archaeologically this makes sense: During most of the Bronze Age, the contacts and cultural evolution of Sicily differed from those on mainland Italy, and indeed there were also contacts with Cyprus and the East Med, in addition to the Aegaean contacts that were also present in mainland Italy. This probably explains the stronger Levantine affinity we see in Sicilians.

Another interesting observation can be made with the Globetrotter analysis in Fig. 3: Namely that in northern Italy, South Italians are clearly the better proxy for the Anatolia_BA-related admixture than Caucasus people, while in central Italy there's no clear preference. This seems to suggest that central Italy was part of the early diffusion area of Anatolia_BA-related admixture, whereas the latter entered the North with the south-to-north population movements within Italy triggered in the Roman era.

Andrzejewski said...

I believe that the increase of mtDNA H from 19% to 40% during the Middle Neolithic times has anything to do with a partial replacement of original LBK-like Anatalian_N/Levant_N with a more CHG-shifted population. Note also that the rate increased from a Middle Eastern frequency of 19% to over 40% in Europe. (It is also speculated that the CHG had BOTH alleles for fair skin SPL instead of just one, as was the case with Anatolian farmers:

Here's an article published by Nature Communications about a study that suggests Europe's modern mtDNA signature was largely established about 6000 years ago, in the mid Neolithic, by people of an unknown origin who largely replaced the early Neolithic farmers, for reasons that aren't yet clear. Although it does also indicate that Bell Beaker folk expanding out of Iberia did have a significant impact during the late Neolithic.

Here's the abstract.

Haplogroup H dominates present-day Western European mitochondrial DNA variability (>40%), yet was less common (~19%) among Early Neolithic farmers (~5450 BC) and virtually absent in Mesolithic hunter-gatherers. Here we investigate this major component of the maternal population history of modern Europeans and sequence 39 complete haplogroup H mitochondrial genomes from ancient human remains. We then compare this ‘real-time’ genetic data with cultural changes taking place between the Early Neolithic (~5450 BC) and Bronze Age (~2200 BC) in Central Europe. Our results reveal that the current diversity and distribution of haplogroup H were largely established by the Mid Neolithic (~4000 BC), but with substantial genetic contributions from subsequent pan-European cultures such as the Bell Beakers expanding out of Iberia in the Late Neolithic (~2800 BC). Dated haplogroup H genomes allow us to reconstruct the recent evolutionary history of haplogroup H and reveal a mutation rate 45% higher than current estimates for human mitochondria.

Andrzejewski said...

One additional interesting point of reference here is that two Non-Indo-European speaking populations have *very* Europoid-looking phenotypes: Mordvins/Mari, who speak Uralic languages and are either EHG/WSHG:

The second population may be a mixture of CHG + Anatolian Farmers, which is the Adyghe/Cherkes who speak a Northwest Caucasus language (postulated by Johanna Nichols to have a phylum with PIE dating back to 12,000YBP (10,000BC):

It is surprising to find out that the putative proximate sources for Yamnaya or other Steppe population speaking Proto-Indo-European languages both look very modern European, despite their non-IE origins.

So, I am attempting to figure out which ancestral group (or groups) is responsible for the modern phenotype of modern Europeans: Eastern Hunter Gatherers (AG3), WSHG (Mari/Komi/Uralics in general) seem to be an admixture of both; Caucasus/Anatolia groups (Adyge are an almost 1:1 ratio admixture of CHG and EEF). It seems to me that both non-Indo-European groups look almost indistinguishable from modern "White" people, but their genetic origins vary far and wide. What I CAN deduct is that perhaps Yamnaya, as an alleged 50:50 admixture of proximate EHG/WSHG/ANE Mari-like groups (original PIE speakers) with a Caucasus (plus Anatolia?) substrate did inherit the "blond hair, blue eyes, fair skin" Nordic-like appearance because all these ancestral groups within the Steppe/Caucasus/Anatalia/Central Asia horizon were basically similar in their appearance?

Matt said...

Looking at their CHROMOPAINTER/NNLS model again and moving some of the bars so total HG / Anatolian contribution is more clear:

Some pattern of WHG donation being elevated in Western European clusters relative to EHG. Ratio differences of WHG:EHG quite different even comparing the "Irish/Scottish" cluster (NWEurope4) to the "Polish/Ukrainian" cluster (EEurope2), let alone "Basque" (WEurope2). It's logical but seems a bit inexplicable why we get this pattern in contrast to the normal qpAdm results we've seen for ages...

@Simon thanks for the info on the patterns you see in the EEMS.

Simon_W said...

Returning to the Etruscan/Lemnian discussion again, I've done some additional reading and, even more so, thinking.

First of all, the Greek name for the Etruscans was "Tyrsenians" or "Tyrrhenians". And the Romans called them "Tusci". But in the Eugubinian tables, written in Umbrian language, we also find the forms "Tursce" and "Turskum". Which demonstrates that these forms are all closely related.

So when Thucydides called the Lemnians "Tyrsenians" he was spot-on. Using the same name for linguistically closely related peoples is nothing but reasonable. However, the controversy starts when he continues to say that these Tyrsenians were "Pelasgians", of the same stock as those who once inhabited Athens. Which means that he believed that they were from Greece. Herodotus didn't call the Lemnians "Tyrsenians", but he recounted that Pelasgians from Attica once had migrated to Lemnos.

Hence it has been theorized that the Tyrsenians/Etruscans were a subsection or tribe of the Pelasgians. It has been suggested that their name may stem from a town called Tyrrha or Tyrsa on the coast of Lydia, the modern-day Tire, where some Pelasgians from Greece may have migrated to.

However, one problem of this theory is, that Homer's Iliad doesn't mention any Tyrsenians, there's not a trace of them in that work. At that time the inhabitants of Lemnos were known as the Sintians. AFAIK the first appearance of the Tyrsenians in Greek literature is in Hesiod's Theogony, dated to around 700 BC. And he clearly locates them in Italy. Later, in a 6th century Homerid hymn, they appear as pirates, catching Dionysus and taking him to distant lands. And Apollonius of Rhodes wrote how the Sintians of Lemnos fled to Santorini, driven out by the Tyrrhenians.

As for Etruscan origins, it's very clear to me that the Villanovan culture of the early Iron Age was already ethnically Etruscan. Because its borders match rather neatly the borders of the historical Etruscans, which cannot be coincidental. Moreover there's an Etruscan inscription from northern Italy found in a Villanovan context. What I did consider was a limited immigration of Pelasgians from Lydia during the late Bronze Age (Recent BA or Final BA), who subsequently acculturated some Italic IEs. After all, there are a few Mycenaean sherds from LBA southern Etruria. But there's one thing making this story rather unlikely: The production of wine and olive oil in Etruria didn't start any earlier than the archaic Etruscan period. But since these goods were very popular in Greece, and produced there since the 3rd millennium BC, it would be strange that Pelasgians from Greece didn't take that knowledge with them. So that basically proves that the Etruscans were not Pelasgians and did not stem from the Aegean.

But why then did Thucydides think that the Lemnian Tyrsenians were Pelasgians from Greece? I don't know, maybe because they had a Hellenized culture like the Etruscans?

Simon_W said...

According to Etruscan tradition, reported by Caecina and Verrius Flaccus, their nucleus appears to have been around Tarquinii, called Tarchna in Etruscan. Because at that place Tarchon, the son or brother of Tyrrhenos, excavated the god Tages from the earth who taught him the essentials of Etruscan religion, and afterwards Tarchon allegedly founded the twelve Etruscan cities. The nail counting system of the Etruscans suggests that this happened aroud 1050 BC. Of course that's just myth, but recent findings in archaeology show that there's some grain of truth in it. I'm particularly thinking of Emma Blake's network analyses. They show that the South Etrurian FBA network near Tarquinii is the most isolated network of FBA Etruria, Umbria and Marche. It's just loosely linked with the network near Vulci, which however shows more ties to Umbria and Marche, and northeastern Etruria is not at all linked to Southern Etruria, but shows connections to Umbria and Marche as well. Which strongly suggests that these latter networks were Italic speaking, not Etruscan. From Southern Etruria the Etruscan language and ethnicity gradually spread through incorporation of formerly Italic communities, until it reached its historical borders.

Check out the three different "factions" found in FBA central Italy:

The rhombus faction near Tarquinii is relatively isolated, the square faction highly integrated, and the circular faction more loosely connected, but over vast distances.

Check out this short lecture by Emma Blake herself, the maps of the networks are very useful:

So in any case my guess for now is that the northern part of the MBA/RBA Apennine culture was Etruscoid, that is, Etruscan-related, even though Bell Beaker and Polada influence had reached there in the Chalcolithic and EBA. Maybe it was like in Sardinia, where we also find Bell Beaker and Polada-related influence, but very low steppe and no IEs before the Romans.

And I guess the MBA/RBA Terramare culture, which arguably goes back to migrants from Hungary, may have been either Proto-Italic or Proto-Sabellic, and its collapse appears to have triggered southwards movements of these people. The Protovillanovan culture either goes back to them, or was influenced by another wave from beyond the Alps. I'm not sure where and when the bifurcation between q-Italic (Latino-Faliscan and Venetic) and p-Italic (Oscan-Umbrian aka Sabellic) has taken place, but maybe the q-Italics were late comers.

Simon_W said...

Moving now on to the present, I found it striking how well the clusters detected by Raveane et al. correspond to modern dialect groups. Here's a map of modern Italian dialectal macro-groups:

SItaly1 = Meridionale estremo
SItaly3 = Meridionale
SCItaly1 = Mediano
NCItaly1/2/3 = Toscano
NItaly1 = Franco-Provenzale
NItaly6 = Ladino e Friulano, most of all
NItaly2/3/4/5 = Galloitalico, most of all

The correspondences are not perfect to be sure, but quite good.
Just strange that the lingua Veneta doesn't get an own cluster. And that the Lombards in contrast get even two of them (probably eastern and western Lombardo).

I do find it remarkable though that Emilians and Ligurians are basically the same in this analysis. Their prehistories were so different! In the Emilia there were Etruscans, followed by Gauls, in Liguria there were Ligurians...

Drago said...

@ Simon_W
That was a very interesting summary & view.
My two cents - I am intrigued by the ''indigenous'' origin of Etruscans. And here I summarise Ash - Proto-Villanova emerges from BA due to connections which increase at the end of the LBA - Mycenean trade, nearest point region being Palafitte-Terramare nearby; a strong central Mediterranean network Etruria - Nuraghes - Cyprus. After the ''Bronze Age collapses'', many groups collapsed, became less dominant, more insular. Terramarre demises, however, the Etruria-Sardinia-Cyprus conection remained viable, hence Etruscans later emerge as the most advanced culture a few centuries later.

Its hard to know what is meant when ancient writers speak of connections between ancient groups, it could be real kinship trough to recent connections, similarities or just literary tropes. But could the Lemnians be part of the above network ?

Simon_W said...

@ Dragos

But the Mycenaean influence in LBA Etruria and northern Italy was very weak, no comparison to the impressive influence in Southern Italy. There are a couple of sherds, no compelling evidence for direct trade contacts, and especially from the few North Italian Mycenaean sherds many were produced in Southern Italy. In the first place the Protovillanovan is part of the Urnfield sphere and the northern influence/similarities predominate over Mediterranean contacts even in the Villanovan phase. One and maybe the only important LBA/early IA place for Mediterranean trade was Frattesina in the Po valley, which entertained a far-flung trading network to the eastern Mediterranean and Northern Africa. They continued to exist after the demise of the Terramare culture and continued their traditions. And they exerted some infuence especially on the Chiusi-Cetona facies of the Protovillanovan, which in my opinion was rather Sabellic than Etruscan.

Simon_W said...

Quite another observation I made today: I compared the allele frequency PCA from Raveane et al. to the preliminary PCA with three Etruscans, of that famous paper that was never published:

Looks like two of the Etruscans fall roughly into the NItaly6 cluster and the third one could belong to the more extreme individuals of the cluster NItaly1. So all of them belong to some of the (geographically) northernmost clusters of modern Italy, and to the ones with the lowest Anatolia_BA, in spite of being from Etruria! Which suggests that there were serious changes in the population structure during the Roman Age most likely, with population movements going from South to North which shifted the whole population southwards, i.e. made it more Anatolia_BA-rich.

And one of the Etruscans had the Iberian Affinity of the Aosta Valley cluster while the two others have the eastern/Balkan shift of the Friuli-Trentino cluster. Now the question arises: Which is from the substratum and which is associated with those who first spoke the Etruscan language? And here it strikes me that the ancient Raetic language, a relative of Etruscan, was spoken particularly in the mountains of the Trentino, so go figure...

Drago said...

@ Simon

''But the Mycenaean influence in LBA Etruria and northern Italy was very weak, no comparison to the impressive influence in Southern Italy. There are a couple of sherds, no compelling evidence for direct trade contacts, and especially from the few North Italian Mycenaean sherds many were produced in Southern Italy. In the first place the Protovillanovan is part of the Urnfield sphere and the northern influence/similarities predominate over Mediterranean contacts even in the Villanovan phase.''

I did not say there was any strong or direct Mycenean influence on Etruria. WE know the Mycenean influences was strongest south, but there was a small indirect impact in Etruria via some central Italian outposts within a Terramarre context.
And there's little debate that the origin of proto-Villanovans is from Urnfield. Somewhere between this, and their central Mediterranean networks, Etruscans and their non-IE language evolved.

''One of the main features that makes Terramare and Protovillanova look intrusive and Hungary-related is their habit of cremating their dead, something the earlier cultures of Italy didn't do. And then there's the fact that the Etruscans preserved the Villanovan tradition best and stuck to the cremation custom longer than other peninsular Italians, longer than the Umbri and Latins for example. ''

Yep, even BB began adopting cremation, via Hungarian infuence.

''So there may be quite a deep relationship between Ligurian and Italic, and in particular Oscan-Umbrian, possibly going back to the Bell Beakers. ''

There are still too many gaps to understand. The Terramarre culture collapse, and much of its Po'ían core zone was depopulated. The rest of mainland Italy was characterised by the Appenine culture, which shows Balkan links in pottery, highland pastoralism and continuation of inhumation burials. Here lies a continuity with Italic, not the north, which after the Terramarre collapse, was under the obviously non-IE Vollanovan group, and later, Celtic migrations.

Simon_W said...

@ Dragos

The Villanovan presence in northern Italy was quite limited, and centered around Bologna and Verruchio. The Etruscan dominance in the Po valley goes back to the late 6th century.

And as for Italic being from the Apennine culture, I found something rather not in agreement with this theory: It's interesting that Oscan-Umbrian shares some features with the Celtic Lepontic:

First of all the *kw>p shift, which might still be a coincidence.
Then they both preserved the word-final *-m, which is an archaism attested among Celtic languages only in Celtiberian.
But then, more convincingly, Oscan-Umbrian also has the *gw>b shift, which is a pan-Celtic feature, not shared by Latin, Faliscan and Venetic.
And most strikingly, Lepontic and Oscan-Umbrian both have *nd>nn and *ks>ss. The same shift also occured in Cisalpine Gaulish, so it might be an areal change. But that's no less important, because it would mean that the ancestor of Oscan-Umbrian once has been geographically close to Lepontic.

On the other hand, Latin shares some innovations with Venetic, at the exclusion of Oscan-Umbrian: *bʰ, *dʰ and *gʰ developed to b, d and g, respectively, in word-internal intervowel position. There are also indications of the developments of *gʷ>w and *gʷʰ>f in Venetic, as in Latin. And of course they both have *kw>kw (q-Italic).

So what does this suggest? There can be no doubt that Lepontic goes back to the Canegrate culture of Northwestern Italy, which was a neighbour of the Terramare culture. Venetic was attested (much later of course) to the northeast of the Terramare culture. And the Protovillanovan culture shows stronger ties to the eastern Urnfield area. So I would say it looks quite reasonable to assume that the ancestor of Oscan-Umbrian has been in the Terramare culture, thereby neighbouring the Lepontic speakers. And somewhere to the northeast have been the ancestors of Latino-Faliscan and Venetic, also in contact with each other. And whereas the decay of the the Terramare culture entailed the migration of some of these people southwards, the movement of some northeastern fringe people speaking languages closer to Latin and Venetic, brought some fresh eastern Urnfield influences across Italy. Granted, apart from Latin and Faliscan there is hardly any linguistic evidence for them in peninsular Italy. This may well be a consequence of the later expansion and southwards spread of Oscan-Umbrian, this is historically attested. According to some Romans though, the language of the Sicels was similar to Latin.

Drago said...

@ Simon
YEs some very interesting connections. I agree that some of these diasporan movements might account for adstrate effects and links you have described, but difficult to see them as formative for Italic as a whole.

Simon_W said...

@ Dragos

So you'd rather say Lepontii and Veneti, or people with an unknown IE language who had acquired the described sound shifts in the Po valley, migrated down the peninsula, thereby exerting an adstrate influence on the Italics who had already been there? Sounds less credible to me...

It's hard to decide where the Italics came from judging from archaeology alone. Archaeology is good at showing us cultural connections and networks, but during prehistory we don't have the direct info what language these cultural groups spoke. With an interdisciplinary approach on the other hand, thanks to linguistics, we can e.g. say which language groups once had contact with each other, and then we can contemplate how this fits onto the networks and cultural connections observed in the archaeological record.

But even just taking the archaeological findings into consideration, the scenario I described makes sense: It's clear that the Protovillanovan culture was multi-ethnic, it encompassed people with various different languages, because it extended from northern Italy and southern Switzerland to the extreme south. But the original impulse must have gone back to just one rather northern/northeastern group which adopted Urnfield fashions first.

The Etruscans go back to the South Etrurian network of the Protovillanovan/FBA. Especially to the Tolfa cluster, which is the southern, more isolated one. Because: Why was it relatively isolated? Because they had a very different language. And this cluster was loosely linked to the Fiora valley cluster in its northern neighbourhood. So there we can see a South Etrurian network arising in the FBA. Most of Etruria at that time however (including the Fiora valley cluster) was connected with Umbria and the Marche, and we can see a large network connecting these regions, with two factions, one of which entertaining a very tight network. This can only be Sabellic, maybe even Umbrian. But with the change from FBA Protovillanovan to early IA Villanovan there was a spread of early Villanovans from the Tyrrhenian coast into the interior of Etruria. That was the time when the Etruscans assimilated part of the Umbri.

We don't see any networks in FBA south-central Italy, including Latium. But the Protovillanovan culture was there as well. And there must have been people speaking dialects related to Latin and maybe many others. But they were still small and dispersed bands with no sense of ethnic unity, and hence there wasn't any linguistic unity either.

Looking at the Protovillanovan archaeological material, people have distinguished two large groups in central Italy: The Tolfa-Allumiere facies on the Tyrrhenian coast south of the river Fiora, and the Chiusi-Cetona facies North of that river in northern/eastern Etruria, Umbria, the Marche and the Romagna. That latter facies had pottery decorations in the Terramare style! And it entertained a connection with Frattesina which continued the Terramare tradition. The material from the Tolfa-Allumiere facies on the other hand resembled the material from eastern Lombardy and the Veneto. (That said, the Chiusi-Cetona facies also shared stuff with the Veneto.) While there was no distinction between Etruria and Latium in FBA1, a split between the two occured with the FBA2, which saw the start of the Latial culture south of the Tiber. And although the cremation rite was preserved longer in Etruria than in Latium, the ancestors of the Latins preserved it until the end of Latial IIA (830 BC), thereby developping a very elaborate ritual and special urns.

Simon_W said...

So, assuming a start of the Protovillanovan around 1150 BC that makes more than 300 years of cremation among the ancestors of the Latins, hardly a short time. The change back to the inhumation spread from the south/east of Italy towards the north/west, and so it may be just a random consequence of geography that it reached Etruria later than Latium.