search this blog

Wednesday, August 24, 2016

On the remarkable genetic homogeneity of Denmark

Open access at Genetics:

Abstract: Denmark has played a substantial role in the history of Northern Europe. Through a nationwide scientific outreach initiative, we collected genetic and anthropometrical data from ~800 high school students and used them to elucidate the genetic makeup of the Danish population, as well as to assess polygenic predictions of phenotypic traits in adolescents. We observed remarkable homogeneity across different geographic regions, although we could still detect weak signals of genetic structure reflecting the history of the country. Denmark presented genomic affinity with primarily neighboring countries with overall resemblance of decreasing weight from Britain, Sweden, Norway, Germany and France. A Polish admixture signal was detected in Zealand and Funen and our date estimates coincided with historical evidence of Wend settlements in the south of Denmark. We also observed considerably diverse demographic histories among Scandinavian countries, with Denmark having the smallest current effective population size compared to Norway and Sweden. Finally, we found that polygenic prediction of self-reported adolescent height in the population was remarkably accurate (R2 = 0.639±0.015). The high homogeneity of the Danish population could render population structure a lesser concern for the upcoming large-scale gene-mapping studies in the country.

Athanasiadis et al., Nationwide Genomic Study in Denmark Reveals Remarkable Population Homogeneity, Genetics Early online August 17, 2016; DOI: 10.1534/genetics.116.189241


andrew said...

Denmark is geographically small, has no major geographic barriers dividing most of its territory, doesn't have a huge total population, and lacks the linguistic and ethnic divisions of countries like Belgium and Finland. Why shouldn't Denmark be remarkably homogeneous?

ArtemisVentus said...

They say that their estimates for height and BMI were very accurate. I would like to know which SNPs they looked at so I could create a simple calculator to parse ancient genomes for these same SNPs and/or use 23andme/ftdna data. Any ideas where I could get the SNP list?

FrankN said...

The following German language PhD study on medieval graves from Usedom (yDNA E1b1b, R1a1a7; mtDNA H,K) also discusses craneometric studies. Interestingly, most Danish series, from MN (Funnelbeaker) through to the Middle Ages, cluster together closely (Abb. 5.5, p. 122), and differ remarkably from other contemporary or ancient Scandinavian and Central Europeans (e.g. Anglosaxons, Icelandic Vikings. The closest relatives to that Danish clade are Nordic BA, Thuringians (early MA Anglians/ Warnians?), and Swedish IA, followed more distantly by Silesian Vandals and Polish CW.

The only exception is the Danish Iron Age series that, far away from the other Danish samples, forms a clade with Swedish MN/LN, Italian Lombards, Crimean Goths and Visigoths.
The study also references aDNA analyses of Danish graveyards dating to the Roman IA (0-400) and the Medieval (1000-1250). According to these analyses, the mtDNA distribution points to a high degree of genetic continuity from the IA until today, However, also it revealed Near Eastern mtDNA R0a in the Roman IA series (stranded Roman legionary? Wife-robbing in Gaul/Britain?). Furthermore, the medieval series had one case of mtDNA U7 hitherto unreported from Scandinavia (Varangian "import"?). "Import" of Roman-mediated East Mediterranean genes might explain why the Danish IA craniometric series falls outside all other (pre-historic) Danish series.

FrankN said...

Corrigendum: On closer inspection of Abb. [Fig.] 5.5 of the a/m study, I have discovered that Danish IA only clusters with Swedish MN/LN (Pitted Ware?), Italian Lombards and Visigoths. Crimean Goths don't fall into that clade but instead form an outgroup to all other series.

Also interesting that Haithabu, technically a Southern Danish series, clusters outside all other Danish series together with Pannonian Lombards (remember our respective E-Mail discussion, Rob?).

Anybody interested in early Slavic history might want to note the tight clustering of Chernyakhov Culture Moldova and Migration Period CSSR. The related Sintana de Mures Culture (Dacians? Ostrogoths?), however, forms part of the Early Medieval "Germanic" (Bajuwarians, Allemani, Franks, Saxons, Anglosaxons etc.) clade.

Shaikorth said...

In fig 5.4 the Danish series also falls outside the main Germanic and West Slavic clusters.

Matt said...

Surprising to me that Globetrotter / finestructure has preference for haplo donation from Britain to model Denmark, rather than Norway/Sweden/Germany. Perhaps Britain is more Anglo-Saxon than I thought? (Or another reason exists for that preference).

One other thing, I think it's still interesting to see the intra recent European PCA, without any other West Eurasian / ancient samples. Those are great, but I do think that they tend to end up dominated by relationships to the other West Eurasian outgroups to Europe, and thereby place populations with similar balance of the ancient components together, while recent intra-Europeans may be more informative for the recent population relationships. Shame there's only really the 8 year old unbalanced PopRes (too light on Eastern Europe) used for most of these... Surely it's not too much of a project for someone to get a balanced and large panel representative across Europe today...

@ andrew

Indeed they do note:
"A recent study of population structure in Great Britain (Leslie et al. 2015) found that the average pairwise FST estimates between 30 geographic regions was 0.0007 – 3.5 times higher than the value we report here (0.0002)."

"To better understand how our FST estimates compare to those from Great Britain, we calculated average FST values in two British population sets using FST values reported elsewhere (Leslie et al. 2015).

The two British sets included regions with geographic distances comparable to those in Denmark.

The first subset comprised of 13 populations from Central/Southern England with a resulting average FST < 0.0002.

The second subset comprised of nine populations from Northern England with a resulting average pairwise FST = 0.0003."

So substructure seems the roughly the same as geographically similar sized areas in England. I bet there may be areas in some parts of Europe that show more structure over distance though (Italy?).

(Though perhaps the Netherlands has more structure "To put this observation in context, a previous study was able to detect subtle population structure along a north-south axis in the Netherlands (Genome of the Netherlands Consortium 2014), a country of almost identical area to Denmark’s.")

@ Davidski, what do you think of this Ohana? "Ohana is a new model-based method that uses the same likelihood admixture model as does STRUCTURE (Pritchard et al. 2000), FRAPPE (Tang et al. 2005) and ADMIXTURE (Alexander et al. 2009), and Newton’s method for optimization. Compared to established methods, Ohana achieves better likelihood estimates (R. Nielsen, personal communication). After running the algorithm, we reported per-country admixture proportions by averaging individual proportions within each country" - Ohana, a tool set for population genetic analyses of admixture components. Interesting is that the software includes a novel function for placing populations in the run on a phylogenetic tree.

The K=4 from their paper of the European populations they use looks to separate nicely, though I think it is not enough K to accommodate BI differences vs others (as they get the "Eastern European" component in defiance of geography).

Shaikorth said...


Have you taken a look at these?

-Europeans with no other W.Eurasians (ancients have no effect as they're projected)

-Europeans again (with much wider sampling), but dimensions constrained by Loschbour. This shows how much of the other PCA's dimension 2 was Basque vs. Maltese recent drift.

Grey said...


"Why shouldn't Denmark be remarkably homogeneous?"

cousin marriage

Denmark is unique in having had *both* the Church's ban on cousin marriage and the lack of geographical barriers physically maintaining population structure

and if you have a population marrying exogamously within an endogamous limit for long enough then eventually they woould end up with everyone equally related to everyone else

i wonder what effect that might have?

(h/t hbdchick)

Matt said...

@ Shaikorth, thanks. Yeah, the former's from Lazaridis 2013, IRC.

I think the latter PCA2 is more what I was thinking about in terms of the distribution of samples (although lacking as much coverage of Central Europe as in PopRes). The separation there between the Northwest European and Eastern European clines relative to the Basque and Sardinian separation is really different compared to Lazaridis PCA, and I'm not 100% sure why... (Possibly the relative amounts of samples). The general shape looks more dominated by recent IBD (though depends on how much % of variance PC1 and PC2 represent).

(PCA2 with some rotation to fit with geography and labeled a bit -

Davidski said...


It seems that the methodology is bloating the British ancestry proportions in the Danes.

Note that the Poles also show a decent chunk of British and French ancestry, which obviously can't be explained by direct gene flow from the UK or even France to Poland, but rather other more plausible stuff, like emigration from northern Germany to Poland during the Middle Ages.

Also, the Germans should be showing a decent level of Polish ancestry. But they're showing high copying of Hungarian haplotypes instead. Again, this looks to be a methodological issue.

At the same time, though, in that diagram above the German and Polish clusters are sister branches in the tree.

However, it's likely that adding more Balto-Slavic populations from Northeastern Europe would rearrange the tree and the entire analysis in a profound way.

Rob said...


Interesting thesis, but I don't have much confidence in older PA arriving at fine-grained details, and aome of the results are odd (eg Yugoslavs clustering with Thurignians, fig 5,4)

As the major Bronze Age questions finish, one might imagine the next phase of research will focus on iron & Roman age remains.

For now, I'm still dismissing PtD and going with the view that Lombards are from Thuringia ;)

FrankN said...

Rob, Shaikorth: Note that Fig. 5.4 in my link relates to early medieval (500-1000), not current populations. The clustering of Visigoths and early South Slavs corresponds to documented Gothic migration routes and the 6th century Heruli settlement around Belgrade. Fitting the Thuringians into that cluster is a bit more difficult, maybe aDNA will help one day.

Rob: going with the view that Lombards are from Thuringia ;)
Forget it! We now have archeological evidence that Tiberius encountered the Lombards at the Stendal-Tangermünde Elbe fording (leaked in a German history forum, publication by the Saxony-Anhalt Archeological Service shouldn't take that long anymore). This confirms the traditional location of the Lombards in the early medieval Bardengau around Bardowick, further up along the Elbe in the Altmark around Stendal, and, in temporary exile due to Roman pressure, in Western Mecklenburg.

Dave: "Note that the Poles also show a decent chunk of British and French ancestry, which obviously can't be explained by direct gene flow from the UK or even France to Poland .."
Two good friends of mine, both with pre-1945 roots in East Pomerania (Slupsk, Gdynia), claim Huguenot descent. Prussia has been a major target of Huguenot emigration, which may have spilled over into protestant parts of today's Poland. Plus, we have the mid-LaTene Celtic migration into Silesia and Lesser Poland, for which archeological studies indicate an origin from around Lorraine.
Otherwise, there has been gene flow from today's Poland to France (Burgundians, Visigoths, possibly Vandals).
The relation between the Lugi on the Oder (Ptolemy's Lugidunum = Legnica?) and those mentionned by the same author in Sutherland is as yet unexplored.

"Also, the Germans should be showing a decent level of Polish ancestry. But they're showing high copying of Hungarian haplotypes instead."
First of all, the "German" sample traditionally used in fact contains two Germans and two Austrians, the latter, IIRC, from the Vienna area. No wonder that sample looks close to Hungarians. Laz 2016 has provided additional German samples, this time from Leipzig, which -surprise, surprise - are a bit more Czech-like. It is high time for regionally differentiated German samples, as we already have for Iberia, Italy or the UK.
However, toponymic research indicates that most of the Slavic settlement indeed originated from the Balkans, and entered today's Germany via Pannonia and Bohemia (the classic route already known from LBK). See, e.g. the following link (I can provide an English language summary of the arguments if there is interest).

The Obotrites in E. Holstein and Mecklenburg, as their name indicates, of course came from the upper/middle Oder. However, recent archeological findings suggest them as late entrants during the 8th, maybe only early 9th century. Possibly, Charlemagne awarded them with formerly Saxon territories for their assistance in the 798 Battle of Bornhöved.
Hence, more of Pannonian and Bohemian than Polish ancestry contribution in Germany might actually fairly represent the path of Slavic Immigration. Note in this repect also the placement of Sorbs in the Laz 2016 PCA. A zoom-in on Mecklenburg, e.g. the R1a hotspot south of Rostock, however, should genetically look quite "Polish", while Vorpommern, especially Rügen, should also be somewhat Danish/ Swedish.

Davidski said...

Poles don't have much, if any, Huguenot ancestry.

Most of the western admixture in Poland comes from the Low Countries and Germany, especially northern Germany, southwestern Germany, and Flanders.

Matt said...

Davidski: Note that the Poles also show a decent chunk of British and French ancestry, which obviously can't be explained by direct gene flow from the UK or even France to Poland, but rather other more plausible stuff, like emigration from northern Germany to Poland during the Middle Ages.

Also, the Germans should be showing a decent level of Polish ancestry. But they're showing high copying of Hungarian haplotypes instead. Again, this looks to be a methodological issue.

Just about this (because I didn't get it at first), if I'm understanding the paper correctly, that the graphic A isn't exactly showing that Poles have a chunk of French ancestry as such, but rather that a minority of French (French nationals living in France) are assigned to the Polish cluster, with essentially all Poles.

That's showing the cluster assignments of the individuals in different European countries, so e.g. the Iberian cluster is about 50:50 Spanish:Portguese with some French individuals. Many of the Hungarian individuals were assigned *to* the German cluster.

After running CHROMOPAINTER and fineSTRUCTURE on the European donor samples, these were organized in the eight major clusters seen in Figure 3A:

Norwegian (NOR), Swedish (SWE), Finnish (FIN), British (BRI), French (FRA), German (GER), Polish (POL) and Iberian (IBE).

There was always one predominant country in the makeup of each of those clusters, with the exception of IBE, in which Spain and Portugal were present at almost equal proportions, and GER, in which samples from Austria, Hungary and the Netherlands were also present at large numbers

(The German cluster looks more like a "Generic Central European" cluster to me, possibly members of the Polish cluster have some more specific drift; I don't find that necessarily implausible for Hungarians to be members of this cluster which is largely German, Austrian, French and Dutch as the Austro-Hungarian Empire would've mixed things up a bit? You might expect to find some people living in Hungary who are actually ancestrally Austrian and that they'd fit in the German cluster.).

(Polish people are known to have emigrated to France and also Britain in the recent past so possibly that is linked to this assignment of some people in the British and French populations to the Polish cluster?)

The analysis where they're actually looking at contributions from the clusters that they assign different population groups into towards another population is limited to the Danes.

Davidski said...


Haha. Oh shit.

Grey said...


"It seems that the methodology is bloating the British ancestry proportions in the Danes."

I may be misunderstanding but isn't another way of looking at the same data that more Danes moved to Britain than France - so they're more related in that way.


"Also, the Germans should be showing a decent level of Polish ancestry. But they're showing high copying of Hungarian haplotypes instead."

Similarly there - lots of German migration to Transylvania in the middle ages.

So if there were
- one point of migration from Hungary to Germany
- two points of migration from Poland to Germany
- two points of migration from Germany to Hungary
then Hungary would top the list?

Rob said...

@ Frank N

"Forget it! We now have archeological evidence that Tiberius encountered the Lombards at the Stendal-Tangermünde Elbe fording (leaked in a German history forum, publication by the Saxony-Anhalt Archeological Service shouldn't take that long anymore). This confirms the traditional location of the Lombards in the early medieval Bardengau around Bardowick, further up along the Elbe in the Altmark around Stendal, and, in temporary exile due to Roman pressure, in Western Mecklenburg."

Confirms what ? Unless Romans erected a statue with inscription saying "hic sunt Langobardi.." it only confirms that the Romans made ventures toward the lower Elbe. Hardly surprising given we already knew that .

The problem with the lower Elbe hypothesis - a long favoured view - is that it is based on circular reasoning and creative interpolation, largely resting on Paul The Deacon's "origin myth', which could be argued to be a fairy-tale, and a plagiarised one at that - having copied Getica, with a few details changed.

Secondly (as I pointed out to you in our personal corrsp.) is the problem with Roman ethnography. That is, it inconsistent with placement of Lombards, and indeed many other "tribes". It is Tacitus to whom the Orthodox perspective cling to, who placed the Lombards at the mouth of the Elbe. But his contemporary (in fact, predecessor) Strabo placed them far to the south, amongst the Suevic nation, just beyond the Hermanduri. Cassio Dio counts the Lombards amongst the various groups raiding Pannonia in the 3rd century. There is then a 200 year gap in mentioned, before they again appear in the middle Danube region.

We then have the archaeological evidence. From c. 450 ADn there appears in the middle Danube a series of inhumed warrior graves in Bohemia and Moravia, which must represent "The Lombards", whose pottery and fibulae forms are undoubtedly Thuringian. By contrast, in the lower Elbe region, one sees an exiguous urn-cremating culture which obviously has nothing to do with the area of historically-attested Lombards.

There is a combined isotopic and archaeological study from Anhalt, which points to only 7% non-local signature. So, whilst not excluding *some* movements from further north, there is no evidence for any large movement of alleged Lombards from, say, lower Saxony. Jaroslav Terjal's physical anthropology analysis of the Pannonian phase of the Lombards points to central Germany.

Finally, one has to piece the political events. From 450 AD, documented archaeologically by the rise of the Niemberger group, the Thuringian realm came into the fore. Their power and raids reached the Danube & Pannonia, as attested by Roman sources lamenting their raid of Aquileia. Finally, the Franks conquered Thuringia c. 530s which is when the Lombards start appearing in Pannonia. What a coincidence. ?

The weight of evidence suggests that the Lombards were some kind of a satellite group/ Cadet branch of Thuringians / refugee group fleeing the Frankish conquest of the Thuringian realm.

Grey said...

"then Hungary would top the list?"

i mean when compared to Poland

Grey said...

lots of directed migrations in the middle ages with various princes and kings importing high tech farming from the general vicinity of Austrasia (h/t hbdchick again)

truth said...

Migrations are not needed to explain. It's plain and simple genetic similarity between close geopgraphic nations.

Matt said...

OT: New paper -

From Whole Genome Sequences Estimate Tibetan highlanders are 6% non-AMH.... Suspect? Quite surprising. (This is presumably inclusive of the pan-Eurasian Neanderthal fraction varying from 5% UP European - 3% present day European and East Asians intermediate). Any access to this one? Ryukendow?

Shaikorth said...


AWood said...

It looks like the Danes made a substantial impact on the British, or alternatively, Danes and ancient British were similar to begin with.

Karl_K said...

"non-AMH sequences in the TIB gene pool (6.17%) was signicantly higher than that in the HAN gene pool (5.86%) (p < 105)(Figures 6A and Figure S27A). Restricting the comparison to Neanderthal-like and Denisovan-like sequences, we did not observe any signicant differences between TIB and HAN as a percentage of Neanderthal-like sequences per individual (Figure 6B), although we did observe some marginally signicant differences whenwerestricted thecomparisontoNeanderthal-specic sequences"

FrankN said...

@Rob, on the Lombards:
1. Half a dozen of Roman authors (V. Paterculus, Strabo, Tacitus, Suetonious, Ptolemy [Laccobardi], C. Dio) record the Lombards as settling on both banks of the lower middle Elbe. All those statements are pretty consistent with each other.
A good overview is provided here
It includes the original Strabo statement, which doesn't say anything about a far southern placement of the Lombards. Strabo takes a Roman perspective, moving from South(-west) to North(-east) (e.g. first Coldui in Bohemia, then Marcomanni, then Lugi). Thus, him listing the Lombards after the Hermunduri means the Lombards lived north of the latter.

2. "Jaroslav Terjal's physical anthropology analysis of the Pannonian phase of the Lombards points to central Germany." And the anthropomorphic clustering from the study I linked above points to Haithabu (Schleswig). If we settle inbetween, where do we land?

3. "The Niemberger Group .." doesn't have much to do with the Thuringians. It was based in Brandenburg, around Berlin, as western outpost of the Odergermanic Groups. The adjacent Elbgermanic culture on the Lower Middle Elbe (i.e. Lombards in my understanding) had strong cultural influence. From the mid 5th century onwards, some Thuringian influence becomes apparent. Re-read the Volkmann paper, esp. Fig. 5

4. "From c. 450 ADn there appears in the middle Danube a series of inhumed warrior graves in Bohemia and Moravia, which must represent "The Lombards" At around the same time, probably a bit earlier, such "Reihengräber" (row graves) appear in Frankish-controlled E. Gaul. Their appearance with the Alemanni is dated to to the mid-5th century, for Thuringia Volkmann (op. cit., p. 11) speaks of "in the course of the 5th century", and in Ostrogoth-controlled Dalmatia they are dated from the late 5th century onwards.
Apparently a fashion that in enormous speed spread across CE, origins of which are obscure. Volkmann links it to Christianisation and Merowingian state-building. More convincing to me is connection to Roman Military cemetaries for (typically Germanic) auxiliaries, as e.g. excavated in Oudenburg/ BE (late 4th/ early 5th cent.). Auxiliary's "Warrior Graves" are already known from 1st. cent. Pannonia (link). Note furthermore that inhumation in row graves was typical for Unetice, and for Santana de Mures, albeit in both cases N-S oriented instead of E-W, as in Frankish/Allemanni/Lombard graveyards. In any case, whatever the origin of those row graves, they don't help much in clarifying the Lombard's geographic origin.


Nirjhar007 said...

Brilliant Karl K , You are the man!...

Chad Rohlfsen said...

Interesting, but I'm going to take non-AMH stuff with a grain of salt until someone else looks at it.

Nirjhar007 said...

This passage is interesting from the research -

Taking advantage of the whole-genome sequence data
that we generated simultaneously in both Tibetans and
Han Chinese, we estimate that Tibetans diverged from
Han Chinese with an average coalescence time of
~15,000–9,000 years. This estimation is much earlier
than 2,750 years ago, estimated by a recent study.4 Our
estimation is less likely to be affected by the archaic sequences
harbored in the whole-genome data, which
could potentially confound the estimation of population
divergence, because the non-AMH sequences were
excluded from the genomic data of both Tibetans and
Han Chinese in this analysis. Therefore, this time estimation
largely reflects the divergence of modern human
ancestry in Tibetans and Han Chinese since the two populations
split from their shared ancestral population.
However, subsequent gene flows from other populations
and between the two populations are expected to influence
the estimation of population divergence, which
we were not able to fully evaluate and control here.
Further efforts are also needed to elucidate the genetic
relationship between Tibetans and Sherpas; for instance,
the reported relationship by recent studies between
Sherpa and Tibetan groups are controversial.56–58 Even
though we observed some degree of differentiation between
the two groups, uncovering their population structure
and inferring their demographic history will require
larger sample sizes.

Anonymous said...

@Rob and FrankN

We recently talked about w.r.t. Frisians on the possibility that some peoples named themselves - or were named and later adopted that name - after geographical area's named after previous peoples.

The Karolingian Bardengau is hard to separate from the Longobards. It's current name in Lünenburger Heide. Lower Elbe region.

FrankN said...

Lombards cont'd:
Here some positive evidence for an early CE Lombard homeland around the Lower Middle Elbe (Bardengau, Wendland, Altmark, W. Mecklenburg):

1. Fürstengräber (Chieftain graves): Unknown to the Jastorf culture, during the early Roman period, "Chieftain graves" make their appearance over a wide area, stretching from SW of Hamburg to east of the Vistula, from Middle Jutland to Moravia (see maps in the link below). As inhumation burials, they are set apart from cremation burials prevailing among Germanics. During the early Roman period, they show a standardised “Lübsow-type” inventory (Lubieszewo, Gryfice County, Pomerania, where 6 such graves were found), including Roman drinking vessels, golden rings, and horseriding accessories.
They are generally believed to have been inspired by East Celtic tradition; similar pre-Roman graves are known from the foreland of the Eastern Alps. The most parsimonious explanation is a Bohemian origin following the Marcomanni conquest. Bohemia, in fact, is among the regions where these graves cluster. The largest clusters, however, are found (a) from SW of Hamburg to Central Mecklenburg, and (b) in Southern Denmark, especially on Fyn and Lolland.
To explain such distribution and clustering, we need to look for a population that participated in the Marcomanni conquest, yet didn't fully leave their original homeland. This fits Roman description of the Lombards, though there may have been other people to which it fits as well. In any case, during the early Roman period we have a new burial habit, most likely reflecting social differentiation, spreading out of Bohemia to the Lower Middle Elbe and beyond (Denmark, Elbe-Weser), but by-passing Thuringia, which speaks against a "Thuringian Homeland" of Lombards at that time.

During the later Roman period, Chieftain graves ("Haßleben-type") are also recorded, with a less standardised gravegood package. They have a somewhat different, and more restricted geographic foci:
- Bohemia (post Marcomannic wars) is lacking,
- the Danish cluster has shifted towards Zealand
- the Lower Middle Elbe cluster has shifted southwards into Elbe-Saale and Thuringia;
- a new cluster arises on the Mecklenburg coast around Rostock.
Several authors have remarked that the distribution of Haßleben-type burials mirrors the -leben (= property, inheritage) toponymy, which is commonly linked to Anglians.

2. Horse burials, more specifically humans inhumed together with horses, are a Lombard trademark and have a/o been reported from Bohemia, Moravia, Pannonia and Lombardy (Collegno). See, e.g.
They weren't a Germanic invention, but a/o quite common for Roman cavallerists, and in the Steppe anyway.

The earliest known Germanic burial of this type stems from Jeetzel in the Hanoveranian Wendland. It has been tentatively dated to the 3rd/4th cent., though the nearby settlement context suggests an even older dating. If placed into the early Roman period, it would connect to the two Chieftain Graves of nearby Hitzacker that both included horseriding accessories as grave goods.
From the Altmark, several horse cremations dating to the younger Roman period are known. Horse burials appear around 450 in the Elbe-Saale Region, around Berlin (Niemberger Group), possibly also Thuringia, and not before 500 in NW Germany (c.f. the following link, p. 76f).
Hence, it looks like the Germanic horse burial tradition originated on the Lower Middle Elbe, from where it spread in all directions, including Britain (Sutton Hoo, cremation), Scandinavia (Viking Age), and, with the Lombards, to Pannonia and Lombardy.

Samuel Andrews said...

Their ADMIXTURE analysis suggests there's recent common ancestry between France/Iberia/Belgium and Hungary/Poland.

Austria doesn't have lots of the Poland/Hungary stuff. Must be because their ancestry is largely from German-speakers who lived West of Austria not long ago.

Shaikorth said...

@Samuel Andrews,

Those populations are likely just not homogenous enough to make their own components, so ADMIXTURE tries to model them as more mixed. IBD shows no elevation in recent common ancestry with Poland compared to their neighbours. Poland doesn't have elevated Iberian IBD compared to Italian IBD or other South Euro IBD either.

Matt said...

It may be that there is less unique drift in those populations, more than there is less homogenity in ancestral populations, and with less drift Ohana does not need to make a new component.

Shaikorth said...


That is a possibility too. By homogenuity I meant within-population IBS sharing, which can vary despite similar ancestral populations, not true differences in population structure.

I found some more European PCA's if you still want to take a look. These don't have Sardinians or Basques, but they have different sample sizes.

Rob said...

@ Frank N

Thanks, some interesting points.

-> About Reihengraber.

Instead of laeti, a more encompassing and flexible view might be to see the RHG to represent the rise of 'big men' or local potentates, as the connection between the Imperial centres with more peripheral regions like northern Gaul, Raetia and Pannonia waned from the mid-late 4th century. Not surprisingly, this began the earliest in Gaul, which showed a marked cultural change in the wake of the 3rd century crisis & an increasingly dis-appeased local aristocracy.

Leaving the appearance of RHG exclusively with Roman military centres is only partly right, but ultimately wrong, because they occur over a wide type of sites, some of which were non-military; but overall, the militarization of the countryside is beyond dispute. Which brings it to the first point above – the withdrawal/ collapse of central Roman authority.

Ultimately, this type of E-W inhumation is just an extension of Roman military burials – as you say, which began in the 4th century. They are not those of laeti, or necessarily Germanic groups. Later, they transformed into representation of mid-level elites in the post –Roman world.

The regions of the Saxons, Lombards and Thuringians saw a different type of inhumation, featuring horse burials, N-S orientation, etc. In Thuringia and sth Bavaria, women & children with artificially –deformed skulls appear. That zone developed under adaptations of certain Hunnic traits – or more generally the Carpathian – Chernyakov zone.

Rob said...

-> Germanic Fürstengräber & your link with of the early Germanic cluster c. Mecklenberg. :

Bemmann’s latest analysis argues that whilst early Roman inhumations evolve from east Celtic traditions, overall they might be more varied, some perhaps resulting from direct early Roman influence.

But the more important aspect is that these early “Lusbow’ type inhumations have no direct link to those from the later part of the Roman Empire.
During the 3rd century crisis, a distinct clustering occurs solely in the middle Elbe-Saale region – lasting only 2 generations, before itself ending. A short lived dynasty.

They appear again in the fourth century, LR period, in West Pomerania, Central
Germany, Northwest Bohemia and Southwest Germany (Alemania). Finally, the impact of the Huns and the spread of RHG come in the mid 5th.

You ultimately conclude : “To explain such distribution and clustering, we need to look for a population that participated in the Marcomanni conquest, yet didn't fully leave their original homeland. This fits Roman description of the Lombards, though there may have been other people to which it fits as well. In any case, during the early Roman period we have a new burial habit, most likely reflecting social differentiation, spreading out of Bohemia to the Lower Middle Elbe and beyond (Denmark, Elbe-Weser), but by-passing Thuringia, which speaks against a "Thuringian Homeland" of Lombards at that time.;

But this is all assumption. First off, we really cannot pinpoint which exact Lusbow graves the early Roman Lombards should match with. But if the ER Lombards really are to be in the mid-lower Elbe region, then some of those could be "Lombards", but what is it actually representing ?

Secondly, again, there appears to be no continuity between the Mecklenberg cluster with later graves, in any region. Rather, what we’re seeing is a shifting pattern of princely graves: from a more diffuse pattern in early Roman times (apart from Rhine-Weser & Frisia), to distinct concentration in the Elbe-Saale & northern Bohemia in the mid –late 3rd, to again spread to cover more areas (the lower Elbe, Alemania, Hevel – Spree & Brandenberg, central Germany) by the 4th. Finally, there a significant cultural & political shifts in the mid-5th mediated via Roman & Pontic-Carpathian events, during & in the wake of the Hunnic collapse.

Thus, no archaeological trail can be seen from Mecklenberg to central Germany to the Danube. In fact, everything points to the contrary. Ultimately, all we are left with is a origin myth written ~ 500 years after the beginning of the said migration, and a couple of toponyms first created a thousand years later, which might not even have anything to do with ‘Lombards’. The only thing we can stipulate is that the Lombards were an Elbe-Germanic group. The only traceable trail is from the middle Danube to northern Bohemia, the latter of which had connections to the middle Elbe-Saale for centuries, emanating from late 3rd century in particular, but most proximate impact was the Thuringian kindgom which began in the late 5th century.

Samuel Andrews said...


My last post was worded in a confusing way. I meant to say the ADMIXTURE analysis shows Poland and Hungary have recent common ancestry and that France, Spain, and Belgium have recent common ancestry, not that they all have recent common ancestry with each other.

Shaikorth said...

Ah, alright. For Poland and Hungary I agree, but IBD doesn't suggest Spain has a recent genetic connection to France or Belgium beyond what can be expected from geography.

Matt said...

@ Shaikorth thanks.

One more I found while browsing was this: (by a Russian blogger I think, don't know who).

There's also a more classic "West Eurasia" PCA by the same blogger:

There are similarities, however, if you align them, whether you use the Georgian-Sardinian position ( to align or the Sardinian-Basque positions (, then it's obvious they aren't quite the same and the relative distances between the European populations are different.

(I think because the more non-European outgroups constrain the European populations to be close together with appropriate shifts, to reflect that their distances to the outgroups are similar).

(All images -

Shaikorth said...

For comparison here's a bunch of IBD-based PCA's.

Finally, this isn't a PCA but a SPA plot which has West Asians and Altaian/Central Asian Turkics on top of Europeans. Might be a bit hard to read but zooming in helps.

With PCA's there's the issue of drifted populations hijacking the end of a dimension (the Falush preprint about ADMIXTURE interpretation also warned about this), perhaps the plot with Loschbour and Stuttgart restricting the eigenvectors could work around this.

FrankN said...

The Reihengräber pattern is quite irritating. On one hand, they seem to have spread eastwards out of Frankish E. Gaul, so authors such as Volkmann link them to the emergence of Merovingian statehood. But we also find them, geographically unconnected, in Ostrogoth-dominated Dalmatia and Slovenia, before they find their way into Bavaria and Moravia.
However, there are two proto-forms that are difficult to connect to direct Roman influence. One is Sintana de Mures - Chernyakov (SdM-C), with N-S orientated inhumation in densely packed, row-like organised graveyards. The other is late 4th/ early 5th cent. graveyards between the Lower Elbe and the Lüneburg heath (Bardengau) that bi-ritually combine urnfields with W-E-oriented row graves (See, e.g. Fig. 14 in the linked Bemmann/ Voß paper for the Issendorf graveyard, 40 km WSW Hamburg, plus corresponding note on Lüneburg-Ochtmissen). Interestingly, along the Lower Elbe, e.g. Otterndorf, there is a zone of mixed orientations (S-N, N-E), while on the Lower Weser (e.g. Sievern, Flögeln), S-N orientation dominates (ibid. p. 20).

So, we seem to have two roots:
a) a Gothic-Dacian one, to the extent SdM-C may be connected to these ethnonyms, as earliest evidence of pure inhumation graveyards in Germanic-influenced populations,
b) Roman Military graveyards, with their E-W orientation and strict adherence to geometry->rows.

The Hunnnic advance results in both traditions getting into contact, and could explain some of the irritating pattern described above (Bemmann/Voß p.22, not reflected in their abstract). It fails, however, to explain the early occurence of E-W oriented row graves south of Hamburg. A closer look at S. Scandinavia, where inhumation dominated in Skane and Zealand (unfortunately ignored by Bemmann) could help in this respect. However, apparently S. Scandinavia already in the Late Roman IA developed the typical ship graves, so it might not qualify as origin of E-W oriented row graves. [However, I now understand why Danish IA craniomorphologically clusters with Swedish MN/LN - sampling was restricted to inhumating Zealand and Bornholm, while from the remainder of Denmark, for using cremation, no samples were available].

The origin of Germanic inhumation burials can be traced to the bi-ritual Wielbark Culture graveyards (men inhumated, women cremated, cf. Benmann/ Voß Fig. 3). From there, bi-ritualism quickly spread westward, reaching E. Mecklenburg (Pasewalk) already during the early Roman Period, and during the later Roman period becoming pre-dominant among Oder Germanics including the Niemberg Group (Volkmann 2013, p. 10f).
Bemmann/Voß see the Wielbark inhumations connected to late La Tene traditions, especially Przeworsk inhumations in Kujawy and Lower Silesia. Volkmann connects them to Bornholm and Zealand. I think an Eastern inspiration, e.g. East Baltic stone graves (Tarands) might also be considered.
SdM-C burials apparently go back to Iazyges, and seem to have been taken over by the incoming Goths.

Rob said...


Final thing; about Lombards, you might wish to refer to "Böhmen in der Spätantike und der
Völkerwanderungszeit unter besonderer Berücksichtigung der Beziehungen zu Baiern und Thüringen", by Jaroslav Jiřík, esp. pp 289 ->

* Also, you mentioned:

"toponymic research indicates that most of the Slavic settlement [of Germany] indeed originated from the Balkans, and entered today's Germany via Pannonia and Bohemia '

I agree, although more specifically, we should look to the Carpathian basin / Slovakia. At this time, there was no south, east & west Slavic; and with emerging languages, isoglossic features split imperfectly- which partly explains South Slavic features in Polabian, or the west Slavic features in Novgorodian

* BTW: from what I've read , RHG begin in Dalmatia only in the final decade of the 5th at the earliest.

FrankN said...

On the Lombards, you might want to check Gregory of Tours (link):
- The Lombard King Alboin married Chlothars daughter Chlotsinda (IV.41)
- The Saxons had entered Italy with the Lombards (IV.42)
- Chlothar and Sigibert had settled the Suevi and other tribes on the land left by Alboins people, but when they returned, the Saxons claimed their land back, though were defeated (V.15)

Apparently, Saxons followed Lombard leadership. How does all that, plus Alboin's marriage to Chlothars daughter, go along with your theory of a Thuringian origin of the Lombards, given that Chlothar with Saxon assistance conquered Thuringia? And, where do you think Saxons could have come into contact with Lombards, so they could decide to follow their king?

Volkmann 2013 (p.14) regards the "Suevi and other people" settled by Chlothar and Sigibert in former Lombard/ Saxon lands as having originated from the Havelland, and places their relocation into the context of Frankish-Avar conflicts. The resettled Suebians gave their name to the early medieval "Schwabengau", the "other people are probably represented by the adjacent "Friesenfeld", "Hassegau", "Engilin", "Warinofeld" and, possibly, "Nordthüringgau" (why else the "North", and the apparent attempt to place other groups between it and Thüringen proper). If so, that would essentially be the Elbe-Saale Region between Magdeburg and Naumburg. However, that area may only have fallen under Saxon control after they assisted the Franks in conquering the Thuringian kingdom in 531. According to Widukind of Corvey, the three battles against the Thuringians took place in Ronnenberg (n. Hannover), Ohrum (n. Wolfenbüttel), and Burgscheidungen on the Unstrut (n. Nebra/ Karsdorf).

Are you aware of the following German language paper that derives the ethnonym "T(h)ueringi" from the Gothic "Teruingi", more specifically their heathen factiom that around 375 decided not to cross into Roman lands (those going there afterwards became known as Visigoths)? Lots of interesting points, I can summarise them if reading German is too cumbersome for you.

Rob said...

Marriage is a matter of expediency and pragmatism. It sits rather well with the fact that the pre-Pannonian Lombards homeland was within the orbitof Thuringia - just like much of central barbaricum was - but not actual Thuringians themselves. If one looks at an accurate depiction of Germanic settlements in late 5th C, Europe, and back tracks from where the Lombards invaders could have come from in their ousting of Herules and Rugii, then there really isn't any other options than northern Bohemia or Thuringia, and whilst we can accept that Germans from the lower Elbe moved south and were incorporated amongst newly appearing southern groups like Alemani, Lombards and Baiovarii, there really isn't any direct tribal or Royal continuity to be had . It's just a fiction made up by PtD to aggrandise his patrons.

The Saxon connection to one or other body doesn't mean much. Because unlike Alemani, Franks, or Thuringii, there were many, non-centralised Saxon groups. Their alliegences were non universal and always shifting; as we very well know . Let's also remember that the Lombards had Bulgar contingents, but no one is claiming they came from the Kuban.

Yes thanks; although I'm a name-place -skeptic; a summary of the Tervingi-Thuringii would be interesting , to save my a weeks worth of reading ;)
(NB Kortland tried to argue gothic came down the Danube instead of the Vistula- a position not well received in general).
For this at least- we will be able to tell through aDNA

FrankN said...

Discussion has moved on, which is probably good as my private Lombard-related dialogue with Rob may not be of interest to everybody. Nevertheless, let me dump a few links here that are pertinent to the original post, i.e. Danish genetics, and at the same time put the Lombards/ Thuringians/ row graves aspects into a wider context, namely Roman Age contacts/ migrations between Rhine and Dnjepr.

1. Early occurence of non-princely inhumation burials on Zealand, plus mid-3rd Cent. Zealand links to Thuringia and Bavarian Suebia (female - exogamy?):

2, The same cultural linkage, this time demonstrated on hairpins. German w. Polish abstract, but the maps should be self-explaining. The analysis towards the end demonstrates that those hairpins are only found in richer graves ("middle class" to "elites"/"queenly")

3. Vellensby graveyard, Bornholm. Late Roman IA, N-S oriented row graves. Cultural relations (female amulet boxes w. medical herbs) to E. Prussia, the Lower Oder, W. Mecklenburg, Thuringia (Wechmar), Lauffen/Neckar, Pannonia and Serbia (Pancsova). Quite some geographic correlation to Haßleben-type Princely Graves.
"Contact and communication between the centres are, among other archaeological finds, reflected in the grave goods within the previously described female burials containing amulet boxes. The objects in their graves indicate the existence of an extensive social network that is, for example, also reflected by the glass finds of the Eggers Type 189, found among the Late Roman Iron Age (C1b) graves of Himlingøje, Nordrup and Crossen. The glass is of Roman origin and had been transported from Cologne to Zealand and Bornholm, and from there onwards to Poland (Kokowski 2004). The distribution of the fibula types Almgren VII 196 and Almgren IX 217 also reflects the existence of trading a network across the Baltic Sea."

4. Connections between Scandinavia and Sintana de Mures. The trail of N- Pontic glassware (Fig. 8) along Dnjestr, Bug, Vistula and through Pomerania into the Danish Isles is easy recognisable as "Gothic" - more surprising are Western extensions (e.g. Gelduba/ Kefeld-Gellep). On reverse connections, e.g. iron combs, maps are lacking. Reference is typically made to NE Germany whatever that means in detail, so the route could have run more westwards.

5. Golden "Kolbenarmringe" (thickened armring with truncated terminal). Golden arm rings are commonly interpreted as royal insignia for their occurence in the grave of Merovingian Childeric I, and literary allusions (Beowulf, Heliand). The paper (Fig.5) shows a 3rd cent. trail from Cernivci/UKR (SdM-C) via Ostrovany/SK, Cottbus/D (Oder-Spree passage!), and Grabow/ W. Mecklenburg to Himlingøje S. of Copenhagen. The latter is commonly regarded as "capital" of Zealand during the 2nd/3rd cent (see, e.g. link 1, and
Fascinating is the later (4-6th cent., Table 1) spreading of these rings to the Merovingian realm, Pannonia, the Lower Don, and even Ireland, plus the fact that several of them bore Roman stamps.

In short - it looks like S. Scandinavia (Zealand/Skane) became a major European political, trading and cultural hub long before Vikings and Varangians entered the stage.

FrankN said...

@Rob: I had a longer summary of the Thueringian-Teringian paper almost finished when IE decided to restart and kill everything I had written. So here only the arguments which I find most convincing:

405/406, the Goth Radagaisus raided Italy with a large heathen following that, according to Zosimus, had been collected between Danube and Rhine. Historians have struggled to identify Radagaisus' Goths. They weren't Visigoths, as those were of Christian (Arianic) faith. Ostrogoths are geographically and politically problematic (they had allied with the Huns, and it was Hunnish auxiliaries driving Radagaisus out of Italy). Hence, Radagaisus' Goths seem to have been a Tervingian faction that had staid north of the Roman borders - possibly representing part of those peoples that acc. to Ammian on their retreat from the Huns had brought unrest to Quadi and Marcomanni.

The paper now notes that the Rada- root is occuring several times in the Thuringian dynasty: Two Radegundes (+512, +587), the (Thuringo-)Varinian king Radigis (c. 550) who claims power over Thuringia as Frankish vasall, and a 7th cent. Thuringian dux Radulf. Apparently, the Thuringian dynasty saw itself as successor to the Gothic (Tervingian) Radagaisus.

The Origo gentis Langobardum notes Agilulf (c. 555-616) as dux Turingis de Taurinus - "Duke of the Pontic Turingians", according to a glosse descended from Anawat (read as "Athana-wald"). Grahn-Hoek inteprets this as reference to Agilulfs descendence from the "original" Teruingians and their rex Athana-ric (+381). (w. background on the political Situation that, according to Grahn-Hoek, lead to the Thuringia-Visigoth solit.)

Finally, Prokopius names Odoaker both as Torcilingorum rex and as rex Gotorum

There is a bit of archeological discussion (those deformed skulls mentionned by you), but essentially the paper delves deep into antique/ early medieval sources. It a/o demonstrates the closeness of Thuringians and Heruli (including the Heruli King Rodulf also bearing a Rada- Name), which implicitly suggests that Heruli and Teruingi might have moved west together.

Based on antique sources, Grahn-Hoek identifies Caucaland, the Tervingians' genitalis terra and first zone of refuge, as the area around the upper Dniestr, i.e. SW Ukraine (the Tervingian's political centre is commonly assumed to have been Moldova and Bessarabia). She suggests that the name Teruingi (in some sources Tyringi) might be related to Tyras, the Dniester's ancient name, i.e. mean "those from the Tyras" (c.f. the Alanic Tanaites on the Tanais = Don).

The article starts with discounting other suggested etymologies. The High Germanic door->T(h)uer sound shift, e.g., ocurred in the 6th cent. at earliest and cannot have transformed the Hermun-duri into Thur-ingians

Rob said...

Thanks Frank, appreciate that

FrankN said...

OT: I was interested whether Caucaland on the Upper Dniestr had anything to do with the Chauci between Ems and Weser. Seems that Chaucan immigration during the late pre-Roman IA is from time to time considered in Romanian literature (see, e.g. link p. 53) but otherwise hasn't been examined in earnest, e.g. by analysing the archeological record yet.

For the Frisians out there: The Chauci are linked to the early medieval Hugmerchi Gau (shire), today's Humsterland NW of Groningen. If correct, this would provide toponymic evidence of a Chauci westward expansion, and early medieval West Frisians being a mix of antique Frisiones and Chauci.

Etymologically, Chauci are being linked to Low German haff, Dutch haf ("lagoon", c.f. German "Hafen"=harbour, haven), Setting them apart from coastal dwellers (Frisiones). Alternatively, connection is being made to OHG houg "mound, hill", c.f. engl. "height", as possible allusion to artificial settlement mounds (terp, Wurt, Wierde, Warft, værft etc.). C.f.

How the Irish Kaukoi (Ptolemy) and Dumfries/ Scotland fit into the picture is yet unclear. To me, current etymologies for Dumfries (1395 "Dumfreiss") as "Fort in the Thicket" (gael.) or "Friar's Hill" (Middle Engl.) are rather unconvincing. The former doesn't make sense for a river port, while the latter would represent a quite unusual combination of Gaelic and Normanised English.

Karen Schousboe said...

I am late to this discussion. But: I just wish to point out that the affinity between the Danish and Swedish cluster is perhaps due to the fact that lage parts of Sweden (and the most populated) were under Danish rule until 166o. I am not sure what role this would play in terms of genetics, though! See