New paper claims to have discovered 17 English Jews killed in 12th-century anti-Semitic attack

September 4, 2022 • 9:20 am

After the Norman conquest of England in 1066, William the Conqueror invited Jews into England. Settling in places like Norwich, York, and Lincoln, they were presumably Ashkenazi Jews, who originated in Europe and the Middle East (my DNA tells me that I’m one of them.)  Their professions were, as a Smithsonian article (second below) recounts, mainly “financiers and moneylenders,” professions that were forbidden for Christians.

But although they were invited in, they weren’t loved, and were ultimately expelled from England in 1290. (Nobody likes the Jews!). History recounts that, around 1190, only 133 years after the Conquest, a group of Jews were killed by Christians who thought they’d do in a bunch of Ashkenazis on their way to the Third Crusade.

Or so the paper below, from a recent issue of Current Biology, tells us. It recounts a genetic analysis of 6 of 17 skeletons—11 adults and 6 children—found in 2004 in a well in Norwich near the location of the city’s ancient Jewish quarter.  This was 18 years ago, and you can see video and photos of mass grave in the “cold case” video below.

The genetic analysis of 6 of these individuals suggests that the skeletons were a group of Ashkenazis, probably ones killed in the 1190 massacre. Click on the screenshot below to go to the Current Biology piece (free access). You can download the pdf here, and I give the reference at the bottom.

First, here’s what was found (photo and caption from 9News.  (You can see a lot more in the video at bottom.)

At least 17 human bodies found at the bottom of a medieval well in Norwich were a group of Ashkenazi Jews who may have been victims of anti-Semitic violence during the 12th century. (NPS Archaeology/BBC)

After the remains were excavated, sorted, and sexed, they found 17 bodies, with 11 adults (a mixture of men and women) and 6 children. The bones were radiocarbon dated to between 1161 and 1216 (95% confidence interval), putting the burial right at the time the Jews of Norwich were killed.

Although a few of the bones were broken, that was from being tossed into the well (head first), which cracked some ribs and vertebrae. As for what killed the individuals, there’s no indication of that: no fractures or indications of blows. They were almost surely killed before being tossed into an empty well, perhaps by stabbing or cutting, or perhaps by smoke inhalation if their houses were burned.

Six of the individuals had DNA extracted from the bones for genetic analysis. The authors didn’t do complete genome sequencing, but did enough to determine that three of them, ranging from 10 years old to young adults, were full siblings—sisters. (You can tell the sex from both the bones and the DNA itself.) These three had identical mitochondrial DNA, showing that they came from the same mother. Another individual was more distantly related to these.

We are also at the stage where we can determine with near certainty the hair and eye color of individuals from DNA, for we know which genes are involved in those traits. As the paper notes,

Two individuals were inferred to have had brown eyes, one with “dark” and one with “light” hair (SB605 and SB676, respectively), while the 0- to 3-year-old boy (SB604) was inferred to have had blue eyes and red hair, the latter of which is associated with historical stereotypes of European Jews.

I wasn’t aware that this is a historical stereotype of European Jews; I would have thought that most of them, like me, had dark hair and eyes, but what do I know from historical stereotypes? And of course two of the three were dark, so what we see is variation, not really a confirmation that these were Ashkenazi Jews. The evidence for that comes from genetic data. First, here’s a reconstruction of the faces of two individuals—a young one and an older one—from both DNA and skulls. These are of course very rough, and don’t tell us much. (I wonder, now that 23andMe has a huge sample of my own genetic data, if they could estimate what I look like from my DNA alone. They already guessed correctly that I have dark hair and eyes.)

The genetic evidence for who these individuals were rests on comparing their DNA with several modern populations from Western Eurasia, as well as looking for sequences of genes that cause disease in modern Ashkenazi Jews. (As an inbred group that probably went through a severe reduction in population size, as well as having substantial intermarriage—”endogamy”—living Ashkenazi Jews have a high frequency of genetic disorders, the most famous being Tay-Sachs disease.

You can see below a plot of where the six individuals whose DNA was analyzed (the black dots) fall in a cluster study—made from what’s known as a principal-components analysis—of the various modern populations. analyzed. Data from living Brits are in the small purple cluster at about 9 o’clock, far away from the Norwich individuals, who fall closer to Southern European and Middle Eastern populations, including modern Ashkenazi Jews, Turkish Jews, and North African Jews. As the authors conclude:

We projected the six Chapelfield genomes on a PCA defined by variation among modern western Eurasian population samples, including modern Jewish individuals. All six Chapelfield [Norwich] individuals project well away from present-day British samples, as well as northern Europeans more generally. Instead, they partially overlap with Southern Europeans, close to Cypriots, modern Ashkenazi, Turkish, and North African Jews. These results are consistent with the Chapelfield individuals having Jewish ancestry (cf. Kopelman et al.)

Click to enlarge:

There’s one more line of evidence that the individuals in the well were Ashkenazim. This is the observation that the six individuals analyzed had a much higher similarity of DNA sequence at the “disease genes” to modern Ashkenazis than to other populations. As the authors note:

To explore this further, we formulated a likelihood function to calculate the exact probability of the six individuals’ observed allele reads at the 159 disease loci, given the allele frequencies of any proposed population.

. . . The maximum likelihood read error rate estimates are notably similar (0.87% and 0.94%, respectively), and crucially these results show that the data are 4,615 times more probable under a model that these individuals were sampled from the modern Ashkenazi population than they were sampled from the modern non-Finnish European population. This approach assumes the six individuals are randomly sampled from either population. Further assessment of the effect of this assumption given that three individuals are siblings suggests that in the case of these data our assumption has a conservative effect on the likelihood ratio

This is scientific jargon saying that, in the end, there’s a much higher likelihood that these genes came from Ashkenazim (modern ones) than from modern European populations. (I’m not sure why they exclude Finns, save that Finns have a high proportion of genes from ancient Siberia.)  The high frequency of alleles that, when homozygous (two copies needed), cause genetic disease, suggest that any “bottleneck” in population size of the Ashkenazi must have occurred before these individuals were killed. That’s several centuries earlier than historians have suggested.

The best guess, then, is that the 17 individuals were Ashkenazi Jews killed by Crusaders who wanted to do in a few Semites on their way to doing in some Muslims in Constantinople. But not every scientist agrees. The Smithsonian piece quotes a dissenter:

Speaking with NatureEran Elhaik, a population, medical and evolutionary geneticist at Lund University in Sweden, casts doubt on the DNA analysis, arguing that the team identified the individuals as Ashkenazi Jews “because that was the only population that they considered.” In response, co-author Ian Barnes, an evolutionary geneticist at the Natural History Museum, tells Nature that local archaeologists and historians know of few other “plausible alternatives” in terms of “other groups that might [have been] in medieval Norwich at the time.”

Given that the paper compared the skeletons in the well with modern populations, most of which were not Jewish, I’m not sure what Dr. Elhaik is on about, but he has published a paper claiming that principal-components analysis is faulty and can be biased to get the results you want. That said, the confluence of the historical and genetic data, and the lack of any other plausible explanation for this slaughter (plus the spot-on estimates from carbon dating to the same period where Jews were being killed in the area), convinces me that the authors are probably right.


Here is an hourlong “Cold Case” video from 2018 about the finding of the bodies, made before any genetic work was done. I haven’t watched the whole thing straight through, but you can see the discovery of the skeletons, other forensic estimation, and people’s best guess at the time what the mass grave told us: 


Brace, S. et al. 2022. Genomes from a medieval mass burial show Ashkenazi-associated hereditary diseases pre-date the 12th century. Curr. Biol.

24 thoughts on “New paper claims to have discovered 17 English Jews killed in 12th-century anti-Semitic attack

  1. Wikipedia has an extensive article on stereotypes surrounding Jews, including on red hair. Today, the prototype of red hair are speckled north-europeans, perhaps, from a place near the arctic. But red hair, more of a fox-brown, and curly hairstyles seem pretty common for Jews, dubbed “Jefro”, according to the article. Further, it cites a paper that suggests endogamy, too, as a reason, as red hair genes are recessive and can accrue more likely in this circumstances.

  2. I’m not sure why they exclude Finns, save that Finns have a high proportion of genes from ancient Siberia.

    Yep, when I got my results back from the longitudinal study All of Us (recommended – it’s free, plus they give you a token remuneration for joining), I learned that I’m 1/8 Finnish, about 3/4 generalized Northern European and the remaining 17% or so Mediterrainean. That told me that my Swedish great-grandparents must’ve each been half-Finnish, or else one was a Finn. Or something like that. (And if it ever became that important to know which, it would actually be possible since they’re both buried right up the hill from me.)

    In any event, if anyone know of a good reference summarizing the distinctive aspects of Finnish vs. N European DNA, I’d be interested.

    1. Surely the problem is they only link to where modern populations that share that DNA live, not where people did live? There are issues with these DNA genealogy website tests, giving different results for the same person from different companies.

      They did not test the DNA of all the bodies. Also, they do not use ancient or historical DNA in this article, do they? I would like to see comparison with Norman-French & Mediaeval English of the same period. Perhaps that would lead to the same results. They ended up burying the remains in the Jewish cemetery.

      As it is impossible to identify how the people died, they latch on to a known incident & connect that with the remains. While they may well be right, it is possible they are wrong. What I thought was most interesting was the suggestion that the Ashkenazi were already genetically constrained at that point, earlier than previously suggested. Endogamy because there were not many potential mates locally, but perhaps also to keep resources within the extended family?

      Sorry to blether on!

      1. High medieval Ashkenazi tradition has it that the Jews from Italy moved to North of the Alps in IIRC the 9th century. This may be where the bottleneck/founder effect of Ashkenazi Jewry was. After that, the population in France/Germany (one interrelated population) was small and endogamous, before the eastern diaspora grew to considerable size. By they way, there were Jewish merchants/settlers in Roman-colonized Germany and probably also France, but it is generally assumed that they didn’t survive the fall of the Roman Empire.

  3. This may be slightly off task, but Joanne Greenberg’s first novel was “The King’s Persons,” a historical novel about the Norwich massacre, which goes into considerable detail about the event. King’s persons refers to the fact that the Jews were not citizens and were therefore property of the king. Indeed, when the king returned from the crusades, he banished the perpetrators to the wilds of Scotland. Greenberg is most famously the author of “I Never Promised You a Rose Garden”; another of her novels was made into a television movie. In all, she has written around a dozen and a half novels, plus short stories.

    1. Richard never returned to England after the crusade. Norwich purchased a charter from the king which was one way he financed the campaign.

    2. Well, no. This book was about the massacre in York, not Norwich. There was no lack of massacres of Jews in England.

  4. Fascinating study. I’ve read about this in several secondary sources over the past few days, including The Forward (

    All the secondary sources I’ve read include the following:

    “I’m delighted and relieved that 12 years after we first started analysing the remains of these individuals, technology has caught up and helped us to understand this historical cold case of who these people were and why we think they were murdered,” Selina Brace, lead author on the paper, said.

    I understand the exuberance of discovery, but think that “delight” is a bit tone deaf. Too bad that this quotation is being picked up and repeated so broadly, as the study is truly remarkable.

  5. The principal component analysis (PCA) presented is meaningless as the first axis explains less that 1% of the variation and the second axis less than 1/2%.

    1. That’s because among West Eurasians (all of the populations are West Eurasian), ca 98-99 % of the variation is within groups, i e purely individual, and only ca 1 percent (depending on what groups you include) is group based variation. That doesn’t mean the PCA is meaningless, it’s a good way to visualize the few group-based differences there are, and the relative distances between the groups tend to reproduce geography.
      If you do a PCA that includes more strongly differentiated populations, East Asians, Mbuti and Papuans, you will get higher proportions “explained” by the PCA .

      1. First, having a priori defined groups in a PCA is a violation of the assumptions. The appropriate model here is discriminant function analysis.

        Second, randomly generated numbers will produce PCAs that explain 1% of the variation.

        1. Are they a priori defined when PCA is an exploratory, not a confirmatory method? And wouldn’t there be a violation of the assumptions in discriminant function analysis with this data set, the data not being independent (more than one individual from one geographical region/presumed group?)
          Current population genetic methods certainly have a lot of possible problems and dangers of producing methodological artefacts. But I find PCA as a visualization tool relatively straightforward, and it’s pretty clear that the two primary factors here are not pure artefacts, though there may be some overfitting.

  6. Hullo, Norwich here!
    I sent Jerry this article earlier in the week…

    The Normans invited the Jews to come to England – that would immediately have established them as hated. In Norwich, Chapelfield was part of the French part if the city. Nothing old survives in this area, but the Jews House as it is known, is 10 minutes walk away in King Street, & us one of the oldest houses in Norwich. I must send PCC[E] photos…

    Norwich was lacking a native saint, so when a boy called William was murdered during the Anarchy, in 1140, & his death was blamed on the Jewish community (the first known example of the “blood libel” as it is known) he was called a saint.

    1. I wonder why they threw the bodies in a well. I always thought it was in the interests of everyone in a community to keep the well water clean.

  7. As a Christian, I am ashamed of so much supposedly perpetrated in the name of Christ, who most certainly was an original peacenik.

    1. Original peacenik? Yeah, especially when you consider the Second Coming! Such a peaceful prophesy upon the return of your “peacenik”. In Luke, he promotes beatings and killings of household slaves as disciplinary measures. There are many other examples that contradict your assertion.

      Also, you’re assuming that a Jesus actually existed; there is no good evidence for a historical Jesus.

  8. Given that archaeologists are told that they can’t even identify the sex of skeletons because they don’t know how the individuals identified, assigning religion to bones is likely to be “problematic” as the kids say nowadays…!

  9. Eran Elhaik seems to be a very controversial researcher. He has theories regarding the origin of European Jews that are not accepted by most scholars.

Leave a Reply