Right Pew, Wrong Church

Do You Have the DNA of Roman-British-Thracian Soldiers in Your Male Line?
Probably Not.


devon uk locationA member of the International Society of Genetic Genealogy (ISOGG) wrote an article online five years ago. Now a substantial number of listers on the discussion board DNA-Genealogy-L believe their male lines may go back to a Balkan legionnaire in Roman Britain. This theory has been enshrined in popular belief, thanks to ISOGG members, who contribute most of the material on Y chromosome DNA to Wikipedia articles.

Read our review from an appendix on Jewish DNA hot spots in England and Wales in our book-in-progress, New Jerusalem:  The Story of Britain’s Earliest Jews and Muslims.
Steven Bird in “Haplogroup E3b1a2 as a Possible Indicator of Settlement in Roman Britain by Soldiers of Balkan Origin,” is, as the title makes clear, most interested in proving a Roman Balkan origin for the haplotype he investigates, now known as Elblbla, the most common type of the haplogroup Elblb (formerly denominated E3b) in Europe. The structure and subclades of this very ancient North African Caucasian lineage have only recently been resolved and overhauled, and the ink is not yet quite dry. But the data used by Bird with the sometimes confused or outdated nomenclature of older reports can still provide valuable clues for our purposes, although one must proceed with caution in making too many differentiations in the tangled branches of the E tree. We must bear in mind that the target haplotype E1b1b1a2 (also called E-V13) represents 85% of the parent haplogroup E1b1b (also denoted as the E-M78 clade) and keep simple E before us without being distracted.

            Bird’s study appeared in one of the first publications of the Journal of Genetic Genealogy, an online journal of the International Society of Genetic Genealogy (ISOGG), founded in 2005 by DNA project administrators of the commercial DNA testing company Family Tree DNA based in Houston, Texas, “who share the common vision of the promotion and education of genetic genealogy.” It is an ambitious work with a very small goal. It uses arguments not only from genetics and statistics but also archeology, geography, history, anthropology and linguistics, often involving such fine points as the epigraphy of a Spanish soldier’s diploma from the British Museum issued in 103 CE and the detailed movements of Thracian cohors II and VII in the Roman army. Where angels fear to thread. Bird’s theory about the origins of Elblb have been enshrined in popular belief. We do not wish to appear ungrateful but there are problems.

Bird’s first mistake occurs in his review of the literature. He misreads Stephen Oppenheimer and represents the author of The Origins of the British as having British E “originating from the Balkan peninsula (26).” If we open Oppenheimer’s book to the page cited (207) we see a map illustrating “Near Eastern [British English for American English ‘Middle Eastern’] Neolithic male migrations via the Mediterranean of E3b [i.e. E1b1b] and J.” The vector standing for the migration of these types launches forth from the Peloponnese in Greece at the cropped lower right corner, obviously intending to suggest origins from that general direction, not “the Balkan peninsula.”  There is no mention of Balkan DNA in Oppenheimer except as part of the bigger picture. The archeological sites Bird adduces as evidence for E settlements in the Bronze Age are not necessarily associated “directly” or solely or chiefly with “proto-Thracian culture,” whatever that term may mean. Nova Zagora in Bulgaria is a Stone Age multi-site. Ezero Culture occupied most of Bulgaria and extended far north into the Danube region of Romania. Yunatsite, Dubene-Sarovka and the other “proto-Thracian culture” examples Bird mentions date to before the Thracians or even the Greeks. They cannot tell us anything about haplogroup E. If anything, all these sites vindicate Oppenheimer’s theory of the demic spread of Middle Eastern (read Anatolian) agriculture, which Bird calls “flawed fundamentally” (27). The center for the diffusion of E in the Balkans is not in Bulgaria or Thrace but northwestern Greece, Albania and Kosovo. The Balkan Peninsula does not have to be the only place from which Bird can manage to derive E and get it to Britain in time to become part of the historical record. It is also strong throughout Greece, Cyprus, the Greek parts of southern Italy, North Africa and even parts of Spain. In fact, its presence in many of those locations is acknowledged to be “due to a founder effect, i.e. the migration of a small group of settlers carrying mostly this lineage (but also a small amount of other North-East African lineages, notably E-M123 and T.” (See www.eupedia.com.)

Despite these failings relating to statement of thesis and validity of arguments, Bird’s work is based on useful data. Three population surveys with frequencies for E in Britain were available to him, the data sets of Capelli, Weale and Sykes. Notwithstanding the nomenclature confusion, only the Sykes data set has true shortcomings, as the Oxford Genetic Atlas Project at the time contained only forty E haplotypes, too small for a valid sample. There are problems comparing them, as Bird realizes, but trends and general conclusions are certainly possible. Before attempting to analyze the haplogroup E variation in Britain, though, we must address the matter of time depth.

We have no quarrel with geneticists’ and genetic genealogists’ methods of gauging coalescence times. Thus, Bird reiterates that the “time to most recent common ancestor” or TMRCA of Cruciani and others led to the “important finding . . . that E-V13 [read 85% of E] and J-M12 [read J] had essentially identical population coalescence times (27).” E and J are companion types that expanded from their Middle Eastern homelands together in the same fashion and probably reinforced each other in multiple phases of gene flow. But who is to say in any specific case of a haplotype that it arrived in Britain 4,000 years ago (TMRCA) or at any subsequent time, including the time when our grandfathers lived. The TMRCA sets a haplotype’s time of origin but not its place of origin, except by inference. We hypothesize that from a host of other factors, chiefly present-day clusters, genetic distance between types and high concentration of haplotype diversity. Using TMRCA, Bird argues that a specific form of E “could not have arrived in Britain during the Neolithic era (6.5-5.5 kya) if it had not yet expanded from the southern Balkans (27).” We prefer to believe that it came to the British Isles at several critical times, first in Neolithic times but later with the Phoenicians, Jews, Egyptians, Iberians and related peoples.

Bird cherry-picks the data to support his Roman Balkan or what might be called Diocletian thesis, but data are data; these are amenable not only to bearing out the general storyline we present but also to supporting, within the same historical context, the existence of certain hot spots for Jewish and Middle Eastern DNA in England and Wales.  We agree somewhat with Bird the Welsh cluster for E is “underestimated by an arbitrary division by Sykes into two geographic regions (‘Wales’ and ‘Northern England’) . . . [creating] an impression of a large number of ‘Eshu’ haplotypes located throughout Northern England, when in fact the Northern English cluster is linked to Welsh cluster geographically (29).” Only, we would see in that Northern English cluster the remains of the historical Welsh Old North (chapters 1 and 7). We would not necessarily see in the Wales-to-Nottingham cluster the fading footprints of “the Ordovices, the Deceangi, the Cornovii, the Brigantes and the Coritani tribes (30),” about whom little is known in any event, but a belt of pre-existing Mediterranean culture reinforced by Roman occupation and somewhat resistant to Anglo-Saxon and Viking intrusions. Another shrinking pocket of the old British culture is shown in the elevated frequencies for both E and J in Strathclyde and Cumbria, part of the Welsh Old North.

Bird has an informative map of Britain illustrating E1b1b distribution according to the Kringing method (34). In this we can trace all the major pockets of Mediterranean and Jewish DNA. Leaving aside Scotland, and aside from the Midlands pocket already mentioned, our eye is drawn to North Wales (along with a clear wall of high incidence surrounding it as though beating back the forces of history on all sides), Dorset, London and East Anglia. It cannot be coincidence that these are the very regions where we have diagnosed the presence of Jews and picked up their trail through the chapters of our book.

As a final note, a 2005 paper by Robert Tarín provides phylogenetic analyses ofE1b1b haplotypes that cast serious doubt on Bird’s assertions and confirm our reading of the evidence. Tarín used 290 individual Y chromosome results to characterize “a separate cluster of mostly Iberian haplotypes which seem to represent a North African entry into Iberia distinct from the E3b [E1b1b] in Europe that may have arisen from Neolithic or other migratory events.” He wrote that “it is unknown whether this finding reflects relatively recent gene flow from the Islamic rule of Spain or an older influx possibly from the Phoenicians”—the same quandary about time frame and coalescence we see above. Utilizing the Y Chromosome Haplotype Reference Database (YHRD), Tarín found levels of the Iberian E haplotype as high as 61% in one Tunisian population (Zriba, near ancient Carthage), while Andalusian Arabs and Tunisian Berbers both showed frequencies of about 7%. We believe this Iberian haplotype is a small, but important Jewish lineage that expanded from Tunisia to the Iberian Peninsula with the Berbers who aided Arab armies in conquering Spain. Interestingly, it accompanied Spanish Jews to Mexico and other places in the diaspora following the events of 1492.  Its distribution in Britain should reveal an implantation originally under the Phoenicians reinforced by periodic migrations of North African and Spanish or French Jews throughout the medieval and early modern periods of British history.


Steven C. Bird, “Haplogroup E3b1a2 as a Possible Indicator of Settlement in Roman Britain by Soldiers of Balkan Origin,” Journal of Genetic Genealogy 3.2 (2007) 26-46.

Robert L. Tarín, “An Iberian Sub-Cluster Is Revealed in a Phylogenetic Tree Analysis of the Y-chromosome E3b [E1b1b] Haplogroup,” published online Nov. 2005 and retrieved Jan. 2012 at garyfelix.tripod.com.

Map shows location of Devon, one possible hotspot for British male haplogroup E.


  Comments: 1

    Your feedback