ashkenazi jews, mediterranean mtdna, mating patterns, and clannishness

here are a couple of thoughts on ashkenazi jews and the apparently high frequencies of mediterranean mtdna found in that population. i was going to include these in my response to prof. macdonald (prolly still will), but since that isn’t happening anytime soon, i thought i’d throw these out there. remember that these are just ideas, so don’t flip out on me!

if it’s correct that 80% of the mtdna of ashkenazi jews is of european — specifically mostly mediterranean, even more specifically very much italian — orgin, then it could very well have been that some male jews (traders?) from judea or alexandria married some roman women, either in or around rome or maybe even in southern gaul. as costa et al. proposed in their paper:

“Overall, it seems that at least 80% of Ashkenazi maternal ancestry is due to the assimilation of mtDNAs indigenous to Europe, most likely through conversion. The phylogenetic nesting patterns suggest that the most frequent of the Ashkenazi mtDNA lineages were assimilated in Western Europe, ~2 ka or slightly earlier. Some in particular, including N1b2, M1a1b, K1a9 and perhaps even the major K1a1b1, point to a north Mediterranean source. It seems likely that the major founders were the result of the earliest and presumably most profound wave of founder effects, from the Mediterranean northwards into central Europe, and that most of the minor founders were assimilated in west/central Europe within the last 1,500 years. The sharing of rarer lineages with Eastern European populations may indicate further assimilation in some cases, but can often be explained by exchange via intermarriage in the reverse direction….

“It is thought that a substantial Jewish community was present in Rome from at least the mid-second century BCE, maintaining links to Jerusalem and numbering 30,000–50,000 by the first half of the first century CE15. By the end of the first millennium CE, Ashkenazi communities were historically visible along the Rhine valley in Germany. After the wave of expulsions in Western Europe during the fifteenth century, they began to disperse once more, into Eastern Europe.

These analyses suggest that the first major wave of assimilation probably took place in Mediterranean Europe, most likely in the Italian peninsula ~2 ka, with substantial further assimilation of minor founders in west/central Europe. There is less evidence for assimilation in Eastern Europe, and almost none for a source in the North Caucasus/Chuvashia, as would be predicted by the Khazar hypothesis, — rather, the results show strong genetic continuities between west and east European Ashkenazi communities, albeit with gradual clines of frequency of founders between east and west….

The age estimates for the European founders might suggest (very tentatively, given the imprecision with present data) that these ancestral Jewish populations harboring haplogroup K and especially N1b2 may have had an origin in the first millennium BCE, rather than in the wake of the destruction of the Jerusalem Temple in 70 CE. In fact, some scholars have argued from historical evidence that the large-scale expansion of Judaism throughout the Mediterranean in the Hellenistic period was primarily the result of proselytism and mass-conversion, especially amongst women.”

just a reminder: the romans were outbreeders. they avoided cousin marriage. the proscriptions against cousin marriage were stronger earlier in the republican period than later, going from no marriage to third cousins or closer to first cousin marriage being allowed by the 200s b.c. but changes to mating patterns and people’s attitudes toward them take time. cosider how long it took for northern europeans to start following the church’s cousin marriage bans in the middle ages — around 300 years in the frankish kingdoms. also, cousin marriage is legal in germany today, and has been (mostly) since the days of luther, but cousin marriage rates remain very low. so i doubt if the roman cousin marriage rates shot up dramatically after the 200s b.c. (although you never know).

what i’m thinking is that the romans might have been quite okay with the idea of marrying their daughters off to some foreigners, especially since they didn’t have a tradition of marrying their cousins. jewish traders from the levant or wherever might not have had a hard time finding a nice roman girl to wed, either in rome itself or even in southern gaul perhaps. this could account for the “major wave of assimilation” in the north mediterranean that costa et al. think that they picked up in their mtdna analysis.

that only “minor founders” came from western/central europe might’ve had to do with the fact that northern europeans like the franks didn’t really start avoiding cousin marriage until the 800s, so they might not have been ready to marry some foreigners at all at that point. and that there was very little introgression whatsoever from eastern europeans should come as no surprise, given that it appears that eastern europeans continued to marry their cousins and be awfully clannish until quite late. when jews moved into eastern europe, they would’ve encountered a populace that barely intermarried between its own members, let alone with some outsiders.

if it’s correct that some jewish blokes married some roman chicks and then we got ashkenazi jews outta that combo AND if the theory of inbreeding/outbreeding/clannishness is right in any way (that’s two big ‘ifs’ there, in case you weren’t counting), then a funny thing to contemplate is that perhaps the jews who had moved northwards into germania in the very early part of the medieval period were some of the least clannish people up in that region, by virtue of the fact that they might’ve had a pretty heavy (outbred) roman ancestry while the northerners had barely begun to outbreed yet. i have to admit that this idea amuses me. (~_^) (just like the thought that much of the european introgression into african-americans likely came from the quite clannish ancestors of our southerners. heh.)

unfortunately, i don’t know what the mating patterns of jews in the levant (or elsewhere) looked like in second or first centuries b.c. not sure that we can guess by biblical proscriptions like those in leviticus either. those were from an earlier time, so it’s not certain they were being followed by our guys in italy. and there was a lot of behavioral variety among jews during this era, too — everything from hellenized jews to pharisees and sadducees, not to mention the jews who thought that jesus and his (universalistic) ideas were pretty cool, so i suspect that there must’ve been all sorts of mating patterns in the middle east at the time. so who knows how clannish (or not) the jews in rome were. no idea.

finally, icymi, i think the subsequent mating patterns of ashkenazi jews went like this. Further Research is RequiredTM.

in any case, mating patterns — and marriage traditions — matter.

see also: Genes Suggest European Women at Root of Ashkenazi Family Tree and Did Modern Jews Originate in Italy?

previously: what did the romans ever do for us? and historic mating patterns of ashkenazi jews

(note: comments do not require an email. roman jewish dude.)


  1. “an origin in the first millennium BCE, rather than in the wake of the destruction of the Jerusalem Temple in 70 CE”: I have read somewhere recently that the great, sudden expulsion/emigration of Jews from Palestine after 70 AD is a myth; it didn’t happen. What happened was a spiritual exile not a literal one. Can anyone direct me to reading on the matter?

    Anyway, if it didn’t happen it’s no surprise that it didn’t lead to mass intermarriage in Rome.


  2. Whether the Romans expelled Jews from Israel or if they merely made life for Jews unattractive there, what we may see is that families migrated within the middle east–and those Jews are to this day indistinguishable from Arabs except to the experienced eye–but young men making their way in the world made for the Jewish center in Rome. With so many single young men arriving taking a Jewish wife was not an option, so they did the best they could. Which turns traditional Jewish law of inheritance on it’s head.

    Roman women may in general have had slim pickings among the Roman men left alive or in country. At any rate, they must have seen that their sisters who were taking the leap were generally doing better than the ones who were not. Women are risk adverse, not risk takers. This was a step up.


  3. A more likely explanation is that large chunks of European DNA are derived from Jewish (or more likely, Israelite) founders. This would be heresy, since Europenas would be horrified to consider that possiblity.

    The most common mtDNA (matrilineal) haplogroup for Jewish women is K at 33⅓%. Second is H averaging to 21%, with Georgian Jews showing the highest concentration of HV at 58%. However half of European women are also H, in addition to H being common among North Africa and the Middle East. Common wisdom explains this away by telling us that Jewish men took H wives who converted into Judaism. This is a little hard to imagine considering the isolation of religious Jews in medieval Europe and the intentional difficulty of female conversion. A more likely explanation is that H was originally Jewish and those women were too rebellious to be good Jewish women and were unusually robust compared to local females. The result would be a constant surreptitious assimilation of these women into the gentile world, thus populating Europe with their mtDNA as more prolific mothers. If this theory is true then it is phenomenal that half of Europe is H, suggesting that an especially successful gene was carried by Jews.

    This theory is supported by the odd fact that pockets of H are geographically coincident with the known patrilineal semitic J haplogroup. Other supporting evidence is that Moroccan Jews show high frequency of H, though slightly lower than European Jews. Libyan, Turkish, Iranian and Indian Jews have a small amount of H haplogroup

    At first I didn’t believe it could be true. But then I asked some Georgian Jews about endogamy among their people and was assured that it was very prevalant mostly due to the difficulty of the terrain. Legend has it that these villages were so insular that they were not affected by the many plagues that swept through the region. A whopping 58% of HV among Georgian Jews suggests that HV was the original Jewish mtDNA haplogroup. The suggestion is that the many mutations of H and V that thrive in Europe came out of a more primitive HV haplogroup that originated in the middle east.

    It seems likely that the high incidence of K among Ashkenazis is a relic of their Khazar origins. There is a reason for this is a result of rabbinical law that allows the son of a prominent rabbi to marry a convert, and thanks to the social resources of rabbis the mtDNA of these converts enjoyed widespread success. This is not terribly obvious until you compare the Y haplogroup of the Kohanim, who were forbidden to marry a convert while being urged to marry the daughter of a kohen. The Ashkenazi kohen haplogroup is practically nonexistent, having been subjected to centuries of negative selection by rabbinical law. In fact an examination of the distribution of the kohen haplotype shows it was much more common in non-Ashkenazi areas where it was associated with matrilineal H.


  4. “perhaps the jews who had moved northwards into germania in the very early part of the medieval period were some of the least clannish people up in that region, by virtue of the fact that they might’ve had a pretty heavy (outbred) roman ancestry while the northerners had barely begun to outbreed yet”

    Your theory about the degree of outbreeding being the driving factor behind ‘clannishness’ is brilliant. But I think this may point to a misunderstanding.

    There is not one suckling animal in the world that evolved to suckle from and be attached to their mother *specifically.* Not one. Plenty evolved to be attached to the first nearby thing that is warm and friendly and feeds them, and the vast majority of the time that is their mother, but if it isn’t, they don’t even begin to give a shit. It can be a cat or a dog or a hairless bipedal ape, if it’s friendly and gives them food in a form they can eat they won’t give much of a shit. They’ll certainly prefer a member of a different species that meets those criteria to their actual biological mother if it doesn’t.

    In the same narrow sense, there is no human who evolved to greatly prefer people who are related to them. All humans evolved to prefer people they already know and people the people they know already know over unrelated people, to varying extents (the variation is what the theory is about, and I agree it’s important!). If you’re part of a large extended family group, then a those folks are related to you. Favoring the ingroup is more genetically favorable than it would be in less related groups, and so the ‘equilibrium ingroup favoritism’ is higher. Not ‘equilibrium familial favoritism,’ *ingroup* favoritism. If you take a baby from a group like that and put it in an entirely different ethnic group, or among the sentient slugs from Betelgeuse, it will still have the same ‘clannish’ predilection. The problem there is to be clannish you need both the predisposition and something like a clan, and few things are more like clans than clans. To get the very most out of the predilection you need an extant clannish clan to be absorbed into. But if the Betelgeuse slugs have a suitable organization, the baby will grow up to fiercely defend Slime Hive Six against all comers. If he’s unfortunate enough to be plopped down in the modern West, he’ll have to content himself with fierce support for the local sports team.

    A child of two people with very high ingroup favoritism should on average have very high ingroup favoritism, regardless of whether those two people are actually from the same clan or ethnic group. He won’t gain as much fitness from his behavior as his ancestors did, but obviously that won’t change his predilections. Only over generations can the reduced fitness of those traits lead to reduced prevalence/potency. At equilibrium, the important stat is degree of relatedness to those that surround you. But we’re not anywhere near equilibrium, IMO. So the most important stat is generations outside of a highly inbred environment. What I quoted above does not automatically indicate that you dispute this, but I think I recall you implying more directly that degree of relatedness = degree of clannishness, even after a single generation.


  5. After the emperor Titus defeated the Jews in Palestine, large numbers of Jewish prisoners were taken back to Rome as slaves. If you’re a slave, who do you marry? Namely another slave. Many native Italian women were slaves in the same households that the Jewish slaves were kept in, so I see this historical intersection as the place where many Jews obtained Italian maternal DNA. Here’s another odd little thought to ponder. After a few generations, many of those slave women would be carrying a lot of genes from Rome’s best families, since Roman men had no qualms about sleeping with their slave women.


  6. Some recent evidence suggests Ashkenazic haplotype K may be Levantine.

    A 2014 study by Fernandez et al have found that Ashkenazi Jews display a frequency of haplogroup K which suggests an ancient Near Eastern origin, stating that this observation clearly contradicts the results of the study led by Richards which suggested a predominantly European origin for the Ashkenazi communities.[61]


  7. If the Jews who moved to Italy were engaged in particular economic niches e.g. goldsmith which were already monopolized by indigenous families then their way into the market may have been through offering large dowries for marriage into those families.



    “In the same narrow sense, there is no human who evolved to greatly prefer people who are related to them.”

    Well that’s the thing. If any aspect of human behavior is modified by whether the object looks like the subject – with self-recognition being a more revealing label for this than kin-recognition imo – then they did.

    That doesn’t mean all traits are like that or even that any traits which are have particularly
    strong effects.

    And if inbreeding increases physical similarity through linkage effects then it will magnify the effect of those traits – if any – that modify behavior according to physical similarity.

    (If the mechanism is visual then they don’t even need to be related just look like they are – hence using tribal tattoos, uniforms etc to trigger the effect artificially.)


  8. Germanic lands (and most others north of Italy) were Christian by the time most of the Ashkenazis’ Jewish ancestors reached them(whereas much of their time in Italy/S.Gaul and the Med. was pre-Christianity). The European fraction in Ashkenazis is mostly Southern perhaps more b/c conversions to Judaism/marriage to Jews were much more taboo in Christian than Pagan societies, thus much rarer. For more tolerant polytheists, Judaism was not necessarily more stigmatized than other exotic cults.


  9. cont.
    The same may also partly explain the greater genetic dilution of Jewish ancestry of (some)Jewish groups in Hindu india and Buddhist China(Kaifeng Jews) than in most Christian and Muslim countries. Iraqi, Persian, and Azerbajani Jews have substantial Levantine ancestry, and most Moroccan and Algerian(though not Tunisian and Libyan) Jews are closer to Sephardim, Ashkenazim, Syrian Jews, and non-Jewish Northern Levantines than to Moroccan Berbers. Yemeni, Ethiopian, and Krypchak Jews may partly(or mostly) derive from the conversion of local gentile polities to Judaism.


  10. correction: “…the conversion of local gentile polities to Judaism, mostly from non-Abrahamic/”Pagan” religions.


  11. Thanks again hbdchick for collecting your thoughts out loud. Since I’ve just been thinking about this very topic myself, I’ll chime in here. :)

    The mtDNA paper is not convincing. All it says is that the mtDNA branches trace to Neolithic Europe. Of course they do! Jews, Italians, and other southern Europeans largely descend from Neolithic farmers. That’s exactly what makes this problem so muddled: any intermarriage that would have occurred would have been between already very similar people. Anyways, we shouldn’t rely much on using present day populations as proxies for older ones because the entire area has changed a lot over the years. So, only ancient DNA can really settle this.

    Thankfully there are some early results. A more recent study, Fernandez et al (2014)[1], samples remains in Syria (8700-6600 BC) and shows basal mtDNA K and generally lots of similarity to early euro farmers. It points to a version of what I and others have been coming to this for a bit now: Ashkenazi being over half European doesn’t make sense. Too many Jewish groups with disparate histories cluster too close together — that was the first clue. And Jews share too much ancestry with disparate Europeans — that was the second. (It’s also important to consider the matrilineal founder effect of 4 women being responsible for the mtDNA of almost half of all ashkenazi jews [2]. That just introduces so much randomness.)

    It looks more like all these populations share sizable descent from some kind of early eastern Med population; more so than previously thought. The DNA clues, combined with what sparse history we know about Jews in and around Italy (generally hostile relations), made be rethink and lower my estimate of european contributions to the Ashkenazi gene pool. Hey, and I started out thinking they were 90%+ euro! ;-)

    Now, there was definitely a decent amount of assimilation wherever Jews settled (it’s important to note that conversions worked both ways.) It’s also possible that structured diversity existed within the Jewish communities from the very beginning, perhaps some of which persisted for a while. Anyways, I’ll wait for a good sampling of ancient DNA. If it comes out looking something like Cypriots do, then we’ll know Ashkenazis are largely an ancient relic. If on the other hand, they look more Bedouin-like, it would tell a different story.

    In summary I think this is a case where the 80/20 rule works better than the 50/50 rule.

    [1] Ancient DNA Analysis of 8000 B.C. Near Eastern Farmers Supports an Early Neolithic Pioneer Maritime Colonization of Mainland Europe through Cyprus and the Aegean Islands (

    [2] The Matrilineal Ancestry of Ashkenazi Jewry: Portrait of a Recent Founder Event (


  12. If Orthodox Judaism dictates one’s Jewishness as being inherited from the matrilineal line, and if the original female founders were Europeans & not Hebrews of the Jewish faith, does this mean that…..nobody’s Jewish?…Oy Vey!!


  13. “It seems likely that the high incidence of K among Ashkenazis is a relic of their Khazar origins. ”

    That makes no sense at all.
    Ashkenazi aren’t Khazars.

    Just like the studies show, they are mainly mixture of Middle Eastern and Mediterranean European ancestry.

    I looked at the Eurogenes K13 Oracle Populations of my maternal Ashkenazi Jewish DNA relative matches.

    My maternal grandmother’s mother Ruth Rosenthal was Ashkenazi Jewish.

    Her father, Max Rosenthal immigrated from Romania,and her mother Irene Hosias immigrated from Latvia.

    What I noticed is that my maternal Ashkenazi Jewish DNA relatives seem to be genetically close to Sicilians,Southern Italians,Greeks, Italian Jews, Sephardic Jews, and Algerian Jews.
    They are not genetically close to Turks, Eastern Europeans,nor Caucasians

    Most of the European of Ashkenazi Jews seems to come from Western Mediterraneans.

    I looked at Ashkenazi Jewish DNA matches with different Mitochondria DNA haplogroups and Y DNA haplogroups.

    Even though there is diversity of haplogroups, the oracle population sharing are similar.

    Top 10 population shared

    MT DNA J1c7a, Y DNA J2

    # Population (source) Distance

    1 Ashkenazi 3.34

    2 East_Sicilian 6.1

    3 West_Sicilian 7.8

    4 Central_Greek 7.86

    5 South_Italian 8.13

    6 Italian_Jewish 8.47

    7 Sephardic_Jewish 8.63

    8 Algerian_Jewish 9.04

    9 Italian_Abruzzo 9.28

    10 Greek_Thessaly 10.7

    MT DNA K1a1b1a, Y DNA E1b1b1c1

    # Population (source) Distance

    1 Ashkenazi 3.52

    2 East_Sicilian 6.47

    3 Central_Greek 6.85

    4 South_Italian 7.38

    5 Italian_Jewish 8.89

    6 West_Sicilian 9.11

    7 Italian_Abruzzo 9.18

    8 Sephardic_Jewish 9.3

    9 Algerian_Jewish 9.9

    10 Greek_Thessaly 10.81

    MT DNA HV5, Y DNA G2b1-M377

    # Population (source) Distance

    1 Ashkenazi 2.8

    2 East_Sicilian 7.83

    3 Central_Greek 8.9

    4 West_Sicilian 9.41

    5 South_Italian 9.65

    6 Italian_Jewish 9.73

    7 Algerian_Jewish 10.2

    8 Greek_Thessaly 10.43

    9 Sephardic_Jewish 10.82

    10 Italian_Abruzzo 10.95

    MT DNA K1a1b1a, Y DNA R1a1a

    # Population (source) Distance

    1 Ashkenazi 5.74

    2 Italian_Jewish 8.59

    3 Sephardic_Jewish 9.81

    4 Algerian_Jewish 10.3

    5 East_Sicilian 10.98

    6 Tunisian_Jewish 11.34

    7 South_Italian 11.68

    8 West_Sicilian 12.09

    9 Libyan_Jewish 12.16

    10 Central_Greek 12.31

    MT DNA K, Y DNA E-L791

    # Population (source) Distance

    1 Ashkenazi 3

    2 East_Sicilian 4.88

    3 Central_Greek 5.76

    4 South_Italian 5.86

    5 Sephardic_Jewish 7.99

    6 Italian_Jewish 8.1

    7 West_Sicilian 8.54

    8 Algerian_Jewish 8.61

    9 Italian_Abruzzo 8.79

    10 Greek_Thessaly 10.68

    MT DNA N1b2, Y DNA J2a1b

    # Population (source) Distance

    1 Ashkenazi 2.47

    2 East_Sicilian 5.75

    3 Central_Greek 7.1

    4 South_Italian 7.15

    5 Italian_Jewish 8.18

    6 West_Sicilian 8.22

    7 Sephardic_Jewish 8.5

    8 Algerian_Jewish 8.86

    9 Italian_Abruzzo 9.16

    10 Greek_Thessaly 10.61

    MT DNA K1a9, Y DNA J1

    # Population (source) Distance

    1 Ashkenazi 6.26

    2 Italian_Jewish 7.19

    3 Sephardic_Jewish 7.94

    4 Algerian_Jewish 9.49

    5 South_Italian 9.92

    6 East_Sicilian 10.23

    7 Tunisian_Jewish 10.68

    8 West_Sicilian 10.96

    9 Central_Greek 11.37

    10 Libyan_Jewish 11.94

    MT DNA N1b2, Y DNA Q1

    # Population (source) Distance

    1 Ashkenazi 4.47

    2 East_Sicilian 6.21

    3 South_Italian 6.37

    4 West_Sicilian 6.43

    5 Italian_Jewish 7.3

    6 Central_Greek 7.45

    7 Sephardic_Jewish 7.61

    8 Italian_Abruzzo 7.75

    9 Algerian_Jewish 8.83

    10 Greek_Thessaly 11.34

    MT DNA L2a1I, Y DNA E1b1b1c

    # Population (source) Distance

    1 Ashkenazi 4.78

    2 Italian_Jewish 6.59

    3 South_Italian 6.7

    4 Sephardic_Jewish 6.88

    5 East_Sicilian 7.05

    6 Algerian_Jewish 7.78

    7 Central_Greek 7.95

    8 Tunisian_Jewish 9.5

    9 West_Sicilian 10.31

    10 Libyan_Jewish 10.6

    MT DNA K2a2a, Y DNA E1b1b1c1

    # Population (source) Distance

    1 Ashkenazi 3.14

    2 East_Sicilian 7.72

    3 Central_Greek 8.49

    4 South_Italian 9.02

    5 Italian_Jewish 9.59

    6 Sephardic_Jewish 9.98

    7 Algerian_Jewish 10.19

    8 West_Sicilian 10.46

    9 Italian_Abruzzo 10.95

    10 Greek_Thessaly 11.33


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s