Other pages & sections of our site:
[Home]  [Y-DNA]   [Contacts   [Groups]  [Haplogroups[Trees]  [Project Blog]  [Special]   [FAQ]
On this page:

Haplogroup R-L21

This page consists of information and musings about the haplogroup known as (shorthand) R-L21 or (phylogenetically) R1b1a2a1a2c. While we are posting it publicly, it is primarily a personal attempt to understand the subject.

R-L21 is highly prevalent in the British Isles and makes up a large part of R-P312, immediately above it on the phylogenetic tree. R-P312 is the haplogroup of about 45% of Taylors and at least three-quarters of the R-P312 Taylors are R-L21.

Note: These statistics are estimated. Fewer than 25% of the 530 R1b (R-M269) Taylor participants have had any SNP testing.
About 70% of Taylor participants are R1b (R-M343), >98% of whom are predicted or confirmed R-M269.


The SNP known as L21 (or M529 or S145) has been called the "Celtic marker". It represents the haplogroup of a majority of the British Isles population prior to Roman Conquest. R-L21 was formed about 4.5 kya and the TMRCA of all R-L21 men lived about that same time.

That L21/M529/S145 is Celtic is disputed by several, who maintain there isn't enough evidence to draw this conclusion and that the Celts were genetically diverse. Indeed, the estimated age of L21 (4.5 kya) suggests it developed on the European Continent and was brought to the islands later.

A problem with the L21=Celtic proposition is that the term was first applied by others (Greeks and Romans) and, until the 18th century, not by the people themselves. They were, in their heyday, a loose grouping of tribes, often at war with each other. They spoke a language in common, with two dialects: Brythonic ("P-Celtic") and Goidelic ("Q-Celtic"); the nicknames refer to the presence of one consonant and the absence of the other.

About 2,400 BC (4.4 kya) Bell Beaker culture artifacts begin to appear in the British Isles. Archeological evidence (hill-forts and mass graves) suggest intense warfare about 400 BC. Ogham inscriptions (a written form of Gaelic) appear in the 5th & 6th centuries AD. See http://www.ancestraljourneys.org/celtictribes.shtml.

It appears there were a number of refugees fleeing to Brittany after the Roman Conquest of Britain.


Within that "loose Celtic grouping", there appear to be (as the color-coding shows) two dozen genetically distinct major groups. Each of those groups also shows sub-divisions in the deep past.

This tree diagram shows how complex are the branches within R-L21.
(Click for larger version.)

More than 1,300 distinct branches of R-L21 have been discovered. (Posted 24 Sep 2016 by Mike Walsh to https://groups.yahoo.com/neo/groups/R1b-L21-Project). The above diagram (also by him) shows only the higher-level branches.

Note: We recommend to those who've tested L21+ join the R1b-L21 project and also its associated Yahoo! discussion group at https://groups.yahoo.com/neo/groups/R1b-L21-Project.

One would suspect connections made between Celtic tribes (as described by Greek or Roman writers) and the phylogenetic tree pictured above. There has been, so far as we know, none made.

Branches of R-L21

Here is a partial list of branches and sub-branches. In all cases, L21+ is assumed

  1. DF63 -- an apparently small group with three known sub-branches and six or seven sub-sub-branches (clades); formed 4.5 kya, TMRCA 4.5 kya
DF13 represents the largest L21 branch. Each listed below is downstream of DF13. Down the left side of the diagram:
  1. DF13 > DF21 -- has five known sub-branches, each having multiple sub-sub-branches, for more than 60 known clades. DF21 formed 4.3 kya, TMRCA 4.3 kya
  2. DF13>DF49 -- has three known sub-branches and 15 clades under them DF49 formed 4.3 kya, TMRCA 4.3 kya
  3. DF13>DF41 -- has nine known sub-branches, with more then 20 known clades.
  4. DF13>FGC11134 -- three known branches, of which the largest is ?146(Y3550/FGC11134?)>CTS4466, with about 47 clades
  5. DF13>FGC5434 -- branches to FGC1191 with six sub branches and 17 clades; formed 4.3 kya, TMRCA 4.3 kya.
  6. DF13>CTS1751 -- a small branch with four known clades; formed 4.3 kya, TMRCA 4.3 kya. .
  7. DF13>516264 ,
  8. DF13>BY573 ,
  9. DF13>CTS119624 ,
  10. DF13>FGC13780    &
  11. DF13>BY2853(?) -- these five, apparently small, branches have few clades under them.
Working down the right side:
  1. DF13>ZZ10>Z253 -- a very large branch with nine sub-branches and more than 70 known clades; equivalents S3218, S471; formed 4.3 kya, TMRCA 4.3 kya. Sometimes called the "Irish Sea" branch.
    BY4064 < BY4090 < BY4087/BY4221 < ZZ5 < Z2534/S868 < Z253/S3218/S471
  2. DF13>ZZ10>Z255 -- three known sub-branches, with 27 known clades; formed 4.3 kya, TMRCA 4.3 kya.
  3. DF13>ZZ10>CTS53386 -- small branch with four known clades.
  4. DF13>ZZ10>MC14 (Y16233) -- small branch with three known clades; formed 4.3 kya, TMRCA 4.3 kya.
  5. DF13>L1335 -- a large branch with two sub-branches, the largest of which has 40+ known clades; formed 4.3 kya, TMRCA 4.3 kya
  6. DF13>L513 -- a very large branch, five sub-branches and more than 70 clades.
  7. DF13>Z251 -- a medium-sized branch, with three known sub-branches and 27 clades; formed 4.3 kya, TMRCA 4.3 kya
  8. DF13>S1026? -- a small branch with three sub-branches and six clades; formed 4.3 kya, TMRCA 4.3 kya
  9. DF13>Z17306?,
  10. DF13>Z15160?,
  11. DF13>Z17904,
  12. DF13>S1051, &
  13. DF13> L371 -- five small branches with no known sub-branches.
Notes: The above information is based on Mike Walsh's tree as of August 2016, which this author takes to be the most robust of those available; it is under constant revision as new discoveries are made. Also, other sources give different versions -- some of which may simply be different choices for SNP nomenclatures. (Many SNPs have multiple names; other SNPs may be different but equivalent.) See:


There are -- or have been -- a number of lively forums for debating issues related to R-L21. Some are

Note: Be wary of information or opinion more than a few years old. The pace of discovery in the most recent two years has been so rapid that much has now been contradicted.
















We conclude that R1b-L21 is "mostly Celtic". What we mean by that is

  1. R1b-L21 was the haplogroup of most of the people who spoke Celtic languages, followed Celtic customs and worshipped Celtic gods when the Romans came to Britain and
  2. A majority of R1b-L21 people at that time were in the British Isles, speaking Celtic languages (Brythonic & Goidelic) following Celtic customs and worshipping Celtic gods. .

We specifically do NOT mean that all Celts were or are L21, nor that L21 is exclusive to Celts. And, this is not to say that there were not some R1b-L21 people in France (Gaul), Spain, Germany and/or Scandinavia. It is reasonable to think that they would have assimilated, adopting languages and customs similar to those of the folks they lived among.

Further, Continental and Insular Europe have been subjected to many wars, climactic changes and disasters. Recent history and current events teach us that such crises create refugees. The Rhine Valley, for example, has been depopulated and repopulated through immigration many times in the historical era.

However, most of the discoveries as to branches are so recent ad the samples so small that little follow-up information has been developed