HAMILTON SURNAME DNA RESULTS AND DISCUSSION

Prepared by Gordon Hamilton, Project Coordinator, Emeritus

1. Introduction and Results Presentation

Most of the analyses reported here were performed by Family Tree DNA (FTDNA) but some participants had their DNA analyzed by DNA Heritage (DNAH, now no longer in existence) or by Ancestry and some of the data was extracted from the Sorenson Molecular Genealogy Foundation (SMGF) database. In the following tables the order of the 111 markers and the conventions used to define each of these markers are those used by FTDNA. Results obtained from other companies were converted to the FTDNA conventions as given in the table at the SMGF web site. According to FTDNA, those marker sites shown in red in the table headings are ones that tend to mutate more rapidly.

The primary data for the sets of results (now nearly 550) currently available are given in 6 tables. These are (1) Group A, (2) Group B, (3) Haplo E, G, J & more I Groups, (4) Haplo R Groups, (5) Group X1, and (6) Group X2. One column in each of these tables, titled Ysearch ID, gives this ID code for most of the participants in order to facilitate examining their marker results in the Ysearch database. Tables (5) and (6) have a column headed Haplo which gives the measured or predicted haplogroup for each participant; the heading for each group in tables (1) to (4) gives the haplogroup information for that group. The meaning of the term haplogroup will be considered later. Modal values for each of the markers for Groups A and B are given in the first row of the first 2 tables. The modal value for each marker was determined by choosing the value that would lead to the actual marker values for all members of the group with the least total number of mutations required.

There are obvious similarities among the results given within each of the groups in tables (1) to (4) as indicated by the color coding. Within each group identical values are colored and any differences due to mutations are highlighted by not being colored. It will be noted that the marker values in each of these groups are quite different from one group to the next so individuals in one of these groups are not closely related through all male lines to individuals in any of the others. However, the similarity of the values obtained within each group strongly implies that the DNA donors in each group have a fairly recent (probably within 10 to 30 generations) common Hamilton ancestor. Since there are 40 different groups in these first four tables, there were presumably 40 different initiating ancestors for these groups, one for each group.

For some time the different groups with matching DNA profiles in the Hamilton project have been named using letters of the alphabet with each subsequent letter being assigned chronologically as each matching group was found. This naming system is now inadequate, firstly because with 40 different groups there are not enough letters of the alphabet to name each one, but also because we now know much more about the haplogroup of each matching group and it is more common to list groups together with related haplogroups. Therefore, the original names of many of the groups have now been changed. Because the original Groups A and B are so large and well known their original names have been more or less retained and the results for these groups are given in two separate tables (Group A and Group B). The names for the other groups are based on their haplogroup families (see later) with numbers being used to distinguish among the different matching groups of these families. The third data table (Haplo E, G, J & more I Groups) has data for matching groups with haplogroups E, G, J and I (other than Groups A and B) while the fourth data table (Haplo R Groups) has data for matching groups in the R haplogroup family. It will be noted that there are 24 matching groups in the R1b family which are thus named R1b-1 to R1b-24.

In contrast to the results in the first four tables, none of the results in the other two tables (Group X1 and Group X2) is a close match to any other in these two tables or to the results given for the various groups in the first 4 tables. The Group X1 table lists unmatched DNA profiles with haplogroups other than those in the haplogroup R1b family while the Group X2 table lists unmatched DNA profiles for those in the haplogroup R1b family. The 40 different initiating ancestors for those in the matching groups (in the first four tables) plus more than 95 required for those in Groups X1 and X2 indicate that there are many different Hamilton lines. As more results are obtained it is expected that matches will be found with more of those in Groups X1 and X2, but it is likely that additional Hamilton lines will also be found. The fact that there appear to be so many Hamilton lines makes DNA analysis particularly useful in sorting out the derivation of lines that do not have an extended paper trail.

In addition to the 6 tables with primary data one other table titled Ancestors is present in this report; it gives the earliest known Hamilton ancestors for each of the participants. The Ancestors table can be accessed either by clicking on this term here or at the top of the page, or by clicking the code for a particular individual in the primary results tables. In that case you will be taken directly to the part of the Ancestors table that has the earliest known ancestor for that individual. With the exceptions that are pointed out in the Ancestors table, many of the participants in the groups given in the first 4 tables were previously unaware of any relationship to each other. Consequently, those interested in these Hamilton lines now have further avenues to explore by traditional genealogical methods.

2. Haplogroups

Since the term 'haplogroup' (often abbreviated as haplo here) is used frequently in the foregoing and subsequent discussion, it seems appropriate to briefly describe here what it means and how it arose. A description of haplogroups can also be found on several internet sites including Wikipedia.

In the DNA analyses summarized in the primary results tables, what is actually being measured at each marker site are so-called STRs or Short Tandem Repeats. The marker values are a measure of how frequently particular sequences of DNA bases are repeated at each marker site on the Y-chromosome. In an attempt to place humans on a phylogenetic tree, anthropologists in recent years have mainly used so-called SNPs or Single Nucleotide Polymorphisms (in other words, substitution of one of the bases on the DNA backbone by a different base). In the SNP analysis investigators assigned capital letters to the various groups that have different initial SNPs. However, since it was found that these initial SNP groups had further SNPs, what one really has is a series of haplogroup families headed by the initial ones characterized by capital letters. In the original attempt to name other members of these families alternating numbers and lower case letters were used to further subdefine the different branches on the human phylogenetic tree. However, it soon became clear that stringing a large number of letters and numbers together to name a subhaplogroup can be very cumbersome so another method has been derived to define each branch in the haplogroup families. By this method the branch is named by starting with the capital letter defining that haplogroup family and then adding the name of the terminal SNP that defines that branch. As an example of these methods for naming haplogroups, those in our Hamilton Group A are currently named by the first method as I1a2a1a1a1 and by the second method as I-L338. In other words, those in Hamilton Group A are in the haplogroup I family but are in a branch of that family defined by a series of SNPs with the terminal one being L338. This second method for naming branches of the haplotree is now being used most often. New SNPs continue to be identified at a fairly rapid pace. This is expected to continue in the immediate future, especially since methods for sequencing all or parts of the Y-chromosome are becoming readily accessable. Therefore, although the initial capital letter used to define a haplogroup family will likely not change, some of the additional numbers and letters to further define the haplogroup may change with time and new terminal SNPs may be found. The best up-to-date listing of haplogroups is maintained by the International Society of Genetic Genealogy at their ISOGG web site. There is frequently a considerable delay before FTDNA takes into account the recent results so their designation of the haplogroup for individual participants is often out of date.

A major difference between mutations at STR sites and those at SNP sites is that the STR sites mutate in a time period of one to a few hundred years (that is one of the reasons why they are so useful in genealogical investigations) while the rate of mutation at SNP sites is measured in the thousand to 10s of thousands of years. Thus, individuals in different haplogroup families could not have a common ancestor along all male lines in the last 10s of thousands of years. Within a haplogroup family individuals on different branches of the family could be related along all male lines in shorter periods of time but still likely only within a few thousand or more years.

By direct SNP analysis the haplogroup of any given Y-DNA sample can be measured. The haplogroup designation given in the heading for each of the groups in the first 4 data tables is the up-to-dare designation given at the ISOGG web site. This haplogroup has been determined or predicted for at least one of the participants in that group. Since the rate of mutation at SNP sites is so slow, all the other participants in that group will have the same haplogroup (thus, making it unnecessary for anyone else in any particular matching group to have a SNP analysis done when at least one participant in that group has had the terminal SNP identified).

The haplogroup for each participant whose results are given in the Group X1 and Group X2 tables is given in the Haplo column of each of these tables. If the haplogroup was determined by direct analysis the result is given in bold green. However, in most cases the haplogroup was not determined by direct analysis because it has been possible to correlate certain patterns of STR marker values with the SNP analysis. Whit Athey has developed a formula for calculating the probable haplogroup from STR data and also at the FTDNA web site most participants receive an estimate of their suspected haplogroup. When predicted these haplogroup designations are given in regular black type in the Haplo column of the Group X1 and Group X2 tables. Since the data in these tables were mainly taken from an FTDNA web site, all haplogroup designations given in these two tables are the current ones given by FTDNA (which may, however, be out of date as indicated in the previous discussion).

As indicated in the tables, the most common haplogroups found for participants in the Hamilton project are various versions of I1 and R1b1a2. However, a few with other haplogroups can be found in some of the tables (except tables for Group A, Group B and Group X2).

3. Group A or I1-1 (Haplo I1a2a1a1a1 or I-L338)

As can be seen from an examination of the data in the Group A table, this group has more members (about 35% of all the participants) than any of the other groups. As indicated at the end of the table there is a group of people with the surname Robertson whose DNA matches the Hamilton Group A profile (more later). In organizing the table attempts were made to keep clusters of similar results together but it should be emphasized that all those in this group are quite closely related with a common ancestor for all of them probably living within the past 500 to 1000 years.

Although some members of this group (see Ancestors) were known to be related, most did not know of any relationship prior to their DNA analysis. Thus, the large number in Group A indicates that this is a widespread group of Hamiltons. Much of the variation in the marker values for Group A occurs at marker sites 570 and CDYa,b. These marker sites are known to mutate quite rapidly so the variation observed is not unusual. One noteworthy difference in the marker values for two participants known to be closely related involves H-054 and H-014. It was initially quite surprising that they differ at 3 marker sites because they are first cousins twice removed; H-054 is the grandson of a first cousin of H-014. However, these three marker sites are known to be on a palindromic or hairpin section of the Y-DNA and the three changes can be explained by only one mutation, a so-called 'Recombinational Loss of Heterozygosity' or RecLOH event. By its nature this RecLOH event had to have occurred in the line of H-014 in the generations from the most recent common ancestor (the grandfather of H-014) of the two participants. Thus, the RecLOH event can be localized to one of only two possible transmission events. The fact that the DNA profile for H-054 is identical to that for a suspected relative (H-232) is further evidence that the mutation has occurred in the line of H-014. Another instance where a RecLOH event seems to have recently occurred in a Group A Hamiton line is in the line to H-123; his known close relatives, H-046 and H-082, do not have this mutation.

One characteristic of the Group A DNA profile is that it is quite unique; very few people other than Hamiltons match this profile even at the 12 marker level. The reason for this is that some of the marker values found for Group A are unusual. This is evident from an early analysis of the dispersion in values for those in haplogroup I1 (referred to as haplogroup I1a in earlier years). Among the first 12 markers for those in haplogroup I1, a value of 13 at DYS385b occurs only 7% of the time, a value of 27 at DYS389-2 occurs only 2.7% of the time, and a value of 12 at DYS439 occurs only 18% of the time. Therefore, it is perhaps not too surprising that the Group A 12 marker signature is limited almost exclusively to those with the Hamilton surname. The corollary, of course, is that if someone with a different surname matches this 12 marker signature he would need to seriously consider the possibility that he is closely related to the Group A Hamiltons in the all male line. This is the situation for several participants whose surname is actually Robertson rather than Hamilton; their results are given at the end of the Group A table. The probability that these Robertsons have a common ancestor with the Hamiltons of Group A is accentuated by the finding that they match so well when many more marker values are compared; the modal values for this group of Robertsons differs from the modal values for the Group A Hamiltons at only 4 markers out of 111. Recent further research has indicated that some Robertson families in Scotland were closely associated with some Hamiltons there, so possibly when surnames came into use two closely related families took different surnames. Several other participants listed in the Group A tables who do not have a known Hamilton ancestor and do not have a Hamilton surname are A-197 (kit 71481), A-350 (kit N28693), B-276 (kit 70799), B-421 (kit 209318), C-491 (kit 282393), D-533 (324694), D-506 (292883), D-314 (150528), F-220 (kit 81909), H-285 (kit 124406), M-161 (kit 39275), M-429 (kit 196476), S-511 (306126), S-159 (kit 52774), S-451 (kit 231420), T-092 (kit 19479), T-479 (kit 264148) and T-490 (kit 282371); their surnames are Adams, Armstrong, Bolton, Bryant, Crawford, Dolmage, Douglas, Douglass, Frost, Harrison, McLain, Marrs, Scott, Smith, Stewart, Thomas and Thompson. The observation that they match so closely this unique Group A DNA profile again strongly implies that these participants share with the Group A Hamiltons or Robertsons a common ancestor along all male lines in the not too distant past.

4. Group B or I1-2 (Haplo I1a3a or I-L1237)

Among the groups that have individuals that match, Group B has the second largest number of participants, about 16% of the total. A few of these were known to be related as indicated in the Ancestors table. The results for two (H-003 and H-017) will be discussed briefly here to emphasize why it is important to obtain results for two or more known relatives. H-003 and H-017 are well documented to be third cousins once removed (separated by nine transmission events) so it is not too surprising that their results are similar. The fact that their results differ at only one site out of 37 markers (and 3 sites out of 67 markers) indicates that there has not been a non-paternal event (unknown adoption, conception out of wedlock, etc.) in the lines of either of these participants back to their known most recent common ancestor (MRCA) who was born in 1763. The observation that their 37 marker profiles differ at one site (DYS570) indicates that there has been a mutation at this site in one of their lines since their MRCA. It seems most likely that the mutation has occurred in the line of H-003 since H-017 has the value (20) at this site most common in Group B. Thus, the 37 marker results of H-017 probably represent those for their MRCA. One could determine whether the mutation has occurred in the line of H-003, and at what generation the mutation occurred, by having closer relatives of H-003 analyzed. Since there is always the possibility of a non-paternal event (estimated to occur about 1 to 5% each generation) or mutation occurring in each person’s lines, it is important to eliminate, or at least become aware of that possibility, by comparing a particular individual's results with those of a known relative. Initially it is best to compare one set of results with those of as distant a relative as possible. If the results are identical then one could conclude that the foregoing possibilities have not occurred in either of the lines. If the results are different then one can zero in on when the mutation or non-paternal event occurred by testing closer relatives.

One interesting aspect of the results for most of those in Group B when compared with the results for those in Group I1-3 (formerly Group C) is that an exact 12 for 12 match in the first 12 markers is found. Therefore, if only the 12 marker test had been run on these samples, one would have concluded that the donors of the samples in Group B were closely related to the donors in Group I1-3. The FTDNA 37, 67 and 111 marker results indicate quite clearly that Groups B and I1-3 are distinct. Such results emphasize the necessity for running at least the 25 marker test, and preferably the 37 marker test, rather than the 12 marker test to make definitive distinctions, especially when the actual values for the markers are fairly common ones for the various marker sites. The values for the first 12 markers in these groups are in fact very common ones as found in the dispersion analysis for I1 individuals referred to previously; the specific value found for each of the first 12 markers for most of those in Groups B and I1-3 is the modal value for that marker in I1 individuals. Another indication that the values for the first 12 markers are common is the observation that many people with other surnames match Groups B and I1-3 exactly at the 12 marker level (currently over 3000 matches in the FTDNA database). Because values for the first 12 markers are the same for Groups B and I1-3, one cannot tell which of these groups H-200 and H-405 should be in. They have been tentatively placed in Group B until results for further markers are obtained.

For each group it is useful to focus on marker values that occur infrequently because their presence in a particular group is what really distinguishes one group from another. As can be seen from an analysis of the dispersion in values for those in haplogroup I1, infrequently occurring marker values that are present in Group B are (frequency in I1 individuals given in parentheses): 7 at DYS459a (4%), 18 at YCAIIa (1.3%), and 14 at GATA A10 (10%). A value of 12 at DYS640 is also quite unique to the Group B Hamiltons. The combination of an 18 at YCAIIa, a 7 at DYS459a and a 12 at DYS640 is so unique to the Group B Hamiltons that other surnamed I1 individuals with these values must seriously consider whether they might be descended from a Hamilton of this type in their all male line. This is the situation for M-183, F-204, A-214 and A-363, F-313, B-324, B-543, J-406, C-424, W-532, and Y-385 whose surnames are Morrison, Frame, Arthurs, Filby, Baker, Bryant, Johnston, Coates, Wormley and Yates respectively; in none of these cases do they know of any Hamilton in their all male lines. However, having the foregoing values along with the excellent match of their other markers to those of the Group B Hamiltons makes it very likely that they have a common ancestor with the Group B Hamiltons in the not too distant past.

Although individuals in Group B have long been given the haplogroup designation I1, about a year ago it was found that this group is positive for SNP Z63 and within this SNP family Group B is also positive for the SNP L1237. Thus, their current haplogroup designation is given as I1a3a or I-L1237.

5. Haplo E, G, J, & more I Groups

In this Table is one small group (E-1) with haplogroup E which is an African haplogroup. It is not surprising the the two participants in group E-1 match because they are known to be related as second cousins. Two other small groups in this table (Groups G-1 and G-2) have the G haplogroup. The two participants in Group G-1 are not known to have a recent common ancestor but the DNA results suggest they do in the not too distant past. Those in Group G-2 are known to be related. Most of the groups in this Table are in the haplogroup I family with seven of them (I1-3 to I1-9) being in the haplogroup I1 family and three others (I2-1, I2-2 and I2-3) being in the haplogroup I2 family. Ken Nordtvedt has estimated that the I1 and I2 subclades separated from one another about 22,000 years ago so there is no possibility that these I2 participants are related through all male lines to the I1 participants within that period of time. As indicated in the Ancestors table, the first five participants in Group I2-1 do, however, share a known common ancestor born about 1650 in Glasgow, Scotland. Those in Group I2-2 also share a known common ancestor but it is not known who is the common ancestor of those in Group I2-3. The final group in this table contains two individuals whose haplogroup is J2; they are known to share the same ancestor.

As indicated in the Table there are far fewer participants in each of the I1 groups in this table than in either Group A or B. For Group I1-3 (formerly Group C) the values of 21,21 or 21,22 at YCAIIa,b are especially notable. It is likely that the 21,21 values at YCAIIa,b arose early in the derivation of this Hamilton line by a RecLOH event (see earlier discussion) because these markers are on a palindrome of the Y-DNA and are thus subject to such events. The DNA profile for the Group C Hamiltons is very similar (including values of 21,21 at YCAIIa,b) to the profiles for several in the Frame and Scruggs surname studies. Such results imply that different members of closely related families with the approximate Group C Hamilton DNA profile took either the Hamilton, Frame or Scruggs surname when surnames came into common use in Lowland Scotland. Individuals in this group have recently tested positive for SNPs Z58 and the terminal SNP L803. Thus, these individuals are currently given the haplogroup designation I1a2a1d1 or I-L803. Recently it has been found that those in Group I1-4 (formerly Group S) also have the Z58 SNP but their currently known terminal SNP is Z138 rather than L803. Thus, the current haplogroup designation for Group I1-4 is I1a2b or I-Z138.

The observation that the 4 participants in Group I1-5 (formerly Group D) match so well is not surprising since they are known to be related. The group is of special interest because they all descend in well documented lines from a grandson (John C. A. Hamilton) of Alexander Hamilton, one of the founding fathers of the US. Participant H-073 is the father of H-039 and grandfather of H-079; H-072 is a third cousin of H-073. As indicated by the results, a mutation has occurred at DYS19/394 in one of the lines from John but it is not known in which line this has occurred. The fact that the results for the third cousins are so similar strongly indicates that there has been no non-paternal event in either of their lines since John C. A., the grandson of Alexander. Furthermore, the results imply that the DNA profile of Alexander, himself, was probably very similar to those in Group D but with the ambiguity that it cannot currently be concluded whether his value for DYS19/394 should be 14 or 15.

The Group I1-6 (formerly Group L) participants represent a somewhat special case because there is some doubt that they are biologically Hamiltons in the all male line. Three of the lines (H-007, H-149 and H-392) can be traced to two sons of a woman with the Hamilton surname whose sons were given the Hamilton surname but whose fathers are unknown. The observation that their DNA profiles match is strong evidence that the same person fathered the two sons. However, since he was presumably not a Hamilton, it is not too surprising that their DNA profile does not match that of any of the other I1 Hamilton participants. Although the other participants (H-174, H-404 and G-395) in Group I1-6 were not aware they were from the same line, their DNA profiles suggests that they are. Recently it has been observed that the Hamilton I1-6 profile matches that of an Osborn profile (Family Group 9). Furthermore, one of the patriarchs of this Osborn line is known to have lived within a mile of the matriarch of these Hamilton lines. It is suspected that this Osborn may have initiated this Hamilton line.

The first two participants in Group I1-7 (formerly Group P) are known to be related; they are from the Earl of Haddington line of Hamiltons so presumably this is the profile for descendants of that line. The third participant in this group is just temporarily placed there until a match with the others is either confirmed or not when results for more markers are obtained. The first two participants listed in Group I1-8 are known to be related to each other and the last two participants in this group are also related. Although the match of one couple with the next is not great, they are close enough that it is likely that the individuals in this group have a common ancestor in the recent past. It is interesting that the results for those in Group I1-8 match quite well the results for several in Group 4 of the Inglis/English DNA project. Presumably, therefore, the Group I1-8 Hamiltons share a common ancestor with this Inglis/English group in the not too distant past.

It is of interest to know how closly or distantly the nine Hamilton I1 groups (I1-1 to I1-9) are related to one another along all male lines. One appraoch to answering that question is to estimate how long ago the subhaplogroups of the various Hamilton I1 groups split from one another. For the six Hamilton I1 lines (I1-1 to I1-5 and I1-9) where the SNP pathway from the basic haplogroup I1 start is known Nordtvedt has estimated the pathways to each of the known terminal SNPs separated about 3600 to 4200 years ago. Therefore, those in one of these I1 haplogroup groups could only have a common ancestor with another along all male lines over 100 generations ago. In other words these six Hamilton I1 groups are not closely related to one another. Since the SNP pathways to the other three Hamilton I1 groups (Groups I1-6 to I1-8) are not known, one does not have a good estimate of when they separated from one another or from Groups I1-1 to I1-5 and I1-9. However, since the DNA profiles for each of these groups differs so much from that of each of the others, one suspects that the separations occurred well over 1000 years ago.

6. Groups with Haplo R1a and R1b1a2

Only two matching groups of haplotypes (Group R1a-1 and R1a-2) have the R1a haplogroup but there are currently 24 different groups (R1b-1 to R1b-24) in our Hamilton project that are in the R1b1a2 haplogroup family. The participants in several of these groups (especially those with few participants) are known to be related as indicated in the Ancestors table but the number of participants in some of the larger groups indicates that these Hamilton lines are fairly major ones. Since the R1b1a2 haplogroup in the human phylogenetic tree is so distant from the I1 haplogroup, there is no possibility that individuals in these two haplogroup families are related to each other through all male lines within the past tens of thousands of years.

The R-L21 haplogroup family is the most common haplogroup family in Scotland so it is not too surprising that many Hamiltons are members of this family which includes at least Groups R1b-5 to R1b-8 and R1b-16. Further terminal SNPs are now known for Groups R1b-5 to R1b-7 but they all still have the L21 SNP. The genetic signatures of those in R1b-5 and R1b-6 are especially close to a DNA profile sometimes referred to as the Scots Modal R1b or the Dalriadic modal since the initiating ancestors of this profile are thought to have been the Dalriadic group who ruled Scotland in the early medieval period. A new terminal SNP for people with this profile has recently been identified. It is L1335 which at least one individual in each of Groups R1b-5 and R1b-6 has. This profile is very common in Highland Scotland, especially among those in Clan Donald and related septs. Participant H-154 in Group R1b-6 has this exact 25 marker profile but most of those in R1b-5 differ from this profile at only one marker, namely DYS449, where R1b-5 has a value of 26 at this site rather than the 30 of the Scots Modal profile. Although it is likely that those in R1b-5 and R1b-6 had a common ancestor in the medieval period, their 67 and 111 marker profiles differ enough that their lines must have diverged at a fairly early time.

The paper trail indicates that participants H-018, H-094, H-031, H-397 and H- 539 in Group R1b-5 are descended from John Hamilton who emigrated from Britain to Concord, MA about 1650 so again it is not too surprising that their results match well. As a result, any direct male Hamilton descendant who suspects he may be derived from this John of Concord can now easily prove or disprove the possibility by having his DNA analyzed and comparing his results to these profiles. In this regard the close similarity of the DNA profiles for P-231, S-327, N-343, F-475 and R-481 to those of the foregoing imply that they may be derived from the same immigrant Hamilton ancestor even though their surnames are not Hamilton. Participant H-084 in Group R1b-5 does not trace his lineage to John of Concord but rather to ancestors in the 1600s in Fife and Argyll, Scotland. The observation that the profile for H-084 is reasonably close to those for the others in Group R1b-5 suggests that they probably all shared a common ancestor in Scotland prior to 1600. The FTDNATiP calculation using 67 markers indicates that there is a 63% chance that H-084 and H-031 share a common ancestor in 12 generations and a 99% chance in 24 generations.

Recently the ancestry of most of those in Group R1b-7 has been clarified with the finding that a well documented descendant (H-293) of the Hamiltons of Mylneholme in the parish of Stonehouse, Lanarkshire has this DNA profile. Since the first 21 participants in this group match so well (a RecLOH event has apparently occurred in the line to H-413 and H-428), it is very likely that all 21 are derived from the Stonehouse line. The four other participants in this group (H-297, H-268, H-355 and H-365) with 37 or more marker values all have ancestors from Bothwell which is close to Stonehouse. These four match each other quite well so they clearly have a recent common ancestor. However, the common ancestor for the Bothwell and Stonehouse groups may be further back in time since the two groups do not match each other as closely although a probable RecLOH event in the Bothwell line may be obscuring how well they match the Stonehouse line. It is likely that the last person (H-310) in this group is also from the Stonehouse or Bothwell line but results for more markers are needed to confirm that. In Mike Walsh's study of haplogroup R-L21 Hamilton Group R1b-7 is placed in his type 1722 with reasonably close matches (within about 1000 years) to Mackenzie, Ross and Sinclair from the highlands.

Two of the participants in Group R1b-9 are known to be related to one another as indicated in the Ancestors table but the similarities of their DNA profiles to those of the others in this group strongly suggests that all have a common ancestor in the recent past. Although the four participants in Group R1b-10 with 37 markers have similar marker values, they differ enough that it is likely their most recent common ancestor was several generations ago, probably before their lines immigrated to America. FTDNATiP calculations indicate that there is about a 50% chance that each pair has a common ancestor in 12 generations and a greater than 90% chance that they share a common ancestor in 24 generations. Paper trails indicate that the five participants in Group R1b-11 are derived from David Hamilton who was born in 1620 in Cambuslang, Scotland. He supported the Royalist cause, was taken as a prisoner by Oliver Cromwell's army after the Scottish army was defeated, and was transported to America in 1651. He settled in Maine and died in 1691 in Berwick, ME. Thus, again any Hamilton who suspects he might be derived from this line can now easily prove or disprove it by having his DNA analyzed.

Those in Group R1b-12 are all known to be related so the observation that they match is not too surprising. There is a quite close match between the results for those in this group with the results for those in Subgroup 2 of the Andrews DNA project so presumably they have a common ancestor in the not too distant past. The four members of Group R1b-13 are all members of Swedish Hamilton lines and are known to be related as indicated in the Ancestors table. The first three participants in Group R1b-14 are known to be related and the similarity of their DNA results to those of the other participants suggests that they have a common ancestor in the recent past even though the fifth participant has a surname (Keyes) other than Hamilton.

7. Group X1 and Group X2

The results for about 18% of the participants are currently placed in these tables because their results do not match those of any of the other participants who have had their DNA analyzed in the Hamilton project. The Group X1 table has unmatched results for those with a haplogroup other than R1b1a2 while the Group X2 table has results for those with a haplogroup of R1b1a2. Even though the Group X1 table has results for participants with a variety of haplogroups, including E (and its sub haplogroups), G2a, G, I1, I1d1, I2a, I2b1, J2 and R1a1a, the Group X2 table has results for a considerably larger number of R1b1a2 participants. This is not too surprising since the R1b1a2 haplogroup is by far the predominant one among British men (see later). It is expected that as more Hamiltons have their DNA analyzed matches with some of the participants whose results are currently in these tables will be found. This has certainly happened to several in the past.

8. Which Groups Are Derived from Well Established and Documented British Lines?

It is well documented that one Hamilton line began in Scotland about 1300 with Walter Fitzgilbert de Hamilton as the patriarch of the line. Walter initially gained prominence in Scotland after the battle of Bannockburn in 1314 when, as constable, he surrendered Bothwell Castle to the victorious Robert Bruce and was rewarded with extensive grants of land in the fertile Clyde valley. Over the years there has been considerable discussion and speculation concerning the ancestry of this Walter with many early commentators suggesting that he was derived from one or more members of the English aristocracy. However, proof for these various speculations has always been lacking. Very recently, Donald Glossinger, a member of our DNA project, has found evidence which appears to identify who were the father and grandfather of this Walter. His research suggests that Walter actually arose from a quite humble background. Michael Stanhope has recently published an article summarizing his research on early Hamiltons in Scotland who may have been ancestors of Walter Fitzgilbert.

As has been discussed in several publications (the best is probably the 1933 book by George Hamilton titled "A History of the House of Hamilton"), many established Hamilton lines in Britain and elsewhere trace their ancestry to Walter Fitzgilbert de Hamilton but a careful reading of the evidence for some of the early connections indicates that many are on tenuous grounds. In any event, the DNA evidence presented here indicates that not all Hamiltons are derived from a single male initiator; only two are required to give rise to about 50% of the participants (those in Groups A and B) but there must have been many additional initiators for those in the other groups. Thus, it is of interest to determine from which ancient and established British Hamilton line each of the various DNA groups determined here are derived from. A start has been made in this direction but many more representatives of long established Hamilton lines need to be analyzed before too many definitive conclusions can be drawn. Such research is continuing.

It has long been accepted that the current Duke of Abercorn is the senior male heir of the Hamiltons with his all male line from Walter Fitzgilbert de Hamilton being very well documented. The DNA of a close relative (H-139) of the Duke has been analyzed and his 37 marker profile differs at only one marker from the modal for Group B. Furthermore, his 67 marker profile differs at only two markers from the modal for Group B. This early result suggested that those in Group B are derived from Walter Fitzgilbert but more recent DNA results have cast doubt on this conclusion. Rather they indicate that there was an early "non-paternal event" (conception by a male other than a Walter Fitzgilbert male line descendant) in the senior male line and that those in Group A rather than those in Group B are the direct all male line descendants of Walter Fitzgilbert. Furthermore, the current DNA results can pinpoint the probable conception where the non-paternal event occurred.

One thing that the current results prove conclusively is that the Sir James Hamilton, 5th of Cadzow, who was born about 1390, married Janet Livingston and died about 1440 had the Group B DNA profile. According to virtually all published genealogies of the Hamilton family, this Sir James Hamilton (subsequently referred to as James1) has been considered to be a gg grandson of Walter Fitzgilbert de Hamilton along the senior male line. One can conclude that James1 has the Group B DNA profile because he is the most recent common ancestor of participants H-139, H-188, H-203, H-230, H-256, and H-360, all of whom have the Group B profile, and all of whom have well documented lines back to James1. Thus, all male line descendants of James1, including those of his son Sir James Hamilton, the first Lord Hamilton (subsequently referred to as James2), would have the Group B DNA profile. James2 is the Hamilton (born about 1415 and died in 1479) who married Princess Mary Stewart, daughter of King James II of Scotland. This marriage brought the Hamiltons close to the throne of Scotland in the 16th century. Since King James VI of Scotland (King James I of England) is a descendant of this union, all Group B Hamiltons can thus claim kinship to the current British royal family and through them to virtually all the royal houses of Europe.

One of the well documented Hamilton lines in Britain is the Raploch line which current evidence (see, for example, the 1933 book “A History of the House of Hamilton” by George Hamilton) indicates was initiated by Walter Hamilton, a brother of James1. The DNA profiles for four well documented descendants of the Raploch Hamiltons (H-190, H-201, H-267 and H-420) have now been determined and they match the Group A profile. Thus, James1 (Group B profile) and Walter, the patriarch of the Raploch line (Group A profile), could only have been half brothers, not full brothers. This indicates that one of the brothers must have been fathered by someone other than a Walter Fitzgilbert all male line descendant, but which one was it?

This question was answered by determining that well documented descendants of lines that branched off from the Walter Fitzgilbert line prior to James1 have the Group A profile. One of these lines is the Preston line; H-217 of Group A is a well documented descendant of this line. The patriarch of the Preston line is Sir John of Hamilton, Lord of Fingalton, born about 1337. Sir John is a grandson of Walter Fitzgilbert de Hamilton, in other words two generations closer to Walter Fitzgilbert than the James1 discussed previously. Another early branching line is the line to the Olivestob Hamiltons; H-103 of Group A is a descendant of this line. The patriarch of the line to the Olivestob Hamiltons is George Hamilton who is a great grandson of Walter Fitzgilbert so he is one generation closer to Walter Fitzgilbert than James1. The only possible explanation for the observation that Preston and Olivestob descendants, as well as the Raploch descendants, all have the Group A profile while James1 and his descendants have the Group B profile is that the break in the senior male Hamilton line occurred in the conception of James1; someone other than a Walter Fitzgilbert all male line descendant was the father of James1.

It is of some interest that the mother of James1 was Janet (Jacoba) Douglas (probably born about 1375), daughter of Sir James Douglas of Dalkeith and his first wife Agnes Dunbar. Agnes apparently died about 1378. By the time of Janet's marriage on 1 November 1388 to Sir John Hamilton, 4th of Cadzow, Sir James Douglas was married to Egidia Stewart, a half sister of King Robert II. Since the Douglas, Dunbar (Agnes was a sister of the Earl of Dunbar) and Stewart families were so prominent in that era in Scotland, the marriage of Janet to Sir John Hamilton of Cadzow undoubtedly brought prestige to the Hamilton family and ultimately led to future generations of the family playing such a prominent role in Scottish society. It may be that Janet knew her first born son was not fathered by her husband so rather than naming him after the Hamiltons she named him after her father. James as a first name was not used by any male line descendants of Walter Fitzgilbert de Hamilton up to that time but in subsequent generations James was used for all first born sons up to the first Duke of Hamilton.

Who was the father of James1? That is not known but one can speculate. If he or other relatives left male line descendants then they should show up with the Group B profile but with a different surname. The Group B DNA profile has unusual STR values at some sites, so unusual in fact that the presence of a combination of these values for a known haplogroup I1 individual virtually ensures that the person must be closely related to the Group B Hamiltons, regardless of surname. These unusual STR values are: 7 at DYS459a, 18 at YCAIIa, and 12 at DYS640. There are several instances of other surnames known to have this Group B profile and who do not have a known connection to a Hamilton. In our Hamilton DNA project are individuals with eight such surnames, including Arthurs (A-214 and A-363) Baker (B-324) Coates (C-424), Frame (F-204), Filby (F-313), Johnston (J-406), Morrison (M-183), Wormley (W-532) and Yates (Y-385). In the Deatherage DNA project there is a line with a very similar DNA profile. Possibly someone with one of these surnames was the father of James1. The possibility that a Wormley might be the progenitor of the Group B Hamiltons has recently been discussed by Nick Wormley at his web site. The very recent finding that L1237 is the terminal SNP for the Group B Hamiltons may make it possible to identify other possibilities for the father of James1; any such person should have the same terminal SNP.

The foregoing analysis suggests that all Hamilton participants in Group B are male line descendants of just one person, namely James1. This is consistent with the limited dispersion of marker values observed for those in Group B; the dispersion is about what would be expected for a 600 year (or about 20 generation) time period. The dispersion of marker values in Group A is greater indicating that the common ancestor for all those in Group A lived in an earlier time period. Also, the fact that there are considerably more participants in Group A than in Group B implies that the Group A line was initiated earlier. It is likely that Walter Fitzgilbert himself is the ancestor for most Hamiltons in Group A but the results would be consistent with some in Group A being derived from earlier male ancestors or male cousins of Walter Fitzgilbert.

(In an article published in December 2011 Henry Lloyd Hamilton has summarized some of the historical events occuring in Scotland around 1390, the time of the conception of James1. These events set a possible scenario for how his father might have been someone other than John Hamilton. As an introduction to this article Gordon Hamilton and Donald Glossinger have briefly summarized the foregoing DNA evidence for a break in the senior male Hamilton line.)

Another ancient Hamilton line is the Earl of Haddington line which is said to be derived from a younger son of Walter Fitzgilbert de Hamilton. However, this conclusion is questionable because two members (H-162 and H-187) of the Earl of Haddington line have now been analyzed in our DNA project and their DNA profiles do not match either those in Group A or in Group B. Instead their profiles are those given in Group I1-7. Since the DNA profiles of H-162 and H-187 match, one can safely conclude that their most recent common ancestor who lived in the early 1700s had a similar profile. Whether earlier generations of the Earl of Haddington line had this profile remains to be determined.

As was discussed previously, there is good evidence that the Mylneholme Hamiltons of Lanarkshire are the ancestors of the R1b-7 Hamilton lines. Group R1b-8 gives the DNA profile of four members of a family of Hamiltons who live in Sweden. They are known to be derived from Captain John Hamilton of Monea and Tullyreny, Co Fermanagh, Northern Ireland, who lived in the late 1600s. Captain John Hamilton is thought to be descended from the Hamiltons of Dalserf, Scotland, which some historians claim connect to the premier male Hamilton line from Walter Fitzgilbert. However, this clearly cannot be correct, at least along an all male line, since these Swedish Hamiltons have a haplogroup (R1b1a2) different from that (I1) of better documented Walter Fitzgilbert descendants.

Another Hamilton line thought to be derived from Walter Fitzgilbert is the line leading to Alexander Hamilton (the first Secretary of the Treasury and one of the founding fathers of the US). As presented previously, the DNA results for his descendants are given in Group I1-5 and they do not match those for any other haplogroup I1 line. Complicating this situation is that there is some lingering doubt whether Alexander Hamilton is biologically really a Hamilton, as has been discussed in a recent biography of the statesman written by Ron Chernow. If his father was James Hamilton, with whom his mother was living when he was born, then the evidence is fairly good that he is derived from one of the well established Hamilton lines. However, as Chernow has pointed out, there have been suggestions that his father was a Stevens, in which case the Group I1-5 results would not be representative of an early Hamilton line. The DNA from some Stevens/Stephens lines has been analyzed and none of those results match Group I1-5 well, although the results for kits 200067 (33/37 versus the modal for Group I1-5) and 220351 (59/67) are just outside the realm usually considered as matches. However, the Stevens line that was in the British West Indies (where Alexander Hamilton was born) around 1750 may not have yet been tested. Thus, at this time we cannot come to any definite conclusions whether Alexander Hamilton's father was, or was not, a Hamilton and whether the DNA results of Group I1-5 represent those of an early Hamilton line.

The observation that the DNA profiles of several ancient and well established Hamilton lines do not match suggests that the Hamilton families who came to prominence in Scotland in the 14th and 15th centuries were not all initiated by one male, but rather were a small number of closely allied families who took the Hamilton surname and who were possibly related through marriage. It is well documented in subsequent generations that a very large number of intermarriages occurred among the various Hamilton lines. Thus, even though the Y-DNA profiles of the various Hamilton lines do not match most of them can be shown to be related to one another through these marriages.

9. Deep Ancestry of the Hamiltons

The haplogroup assigned to each of the DNA groups gives some information about the deep ancestry of that group and where ancestors that lived 10,000 years or more ago may have come from. The R1b haplogroup is the most common (over 50%) among European men. It is found throughout Europe and is especially prevalent among Celtic and Basque populations. Ancestors of those in the R1b haplogroup are believed to have migrated into Western Europe from the east about 7000 to 9000 years ago when the glaciers receded after the last ice age. The common subclade R-L21 in the Hamilton project (Groups R1b-5 to R1b-8 and R1b-16) separated about 4000 years ago. Ancestors of those in haplogroup G are thought to have been Neolithic farmers from the Middle East who were the first to practice agriculture in Europe. This haplogroup is more common in Southern and Eastern Europe but is found at low frequencies in other parts of the continent as well. E is an African haplogroup but it occurs to a small extent in Europe and Britain, possibly arriving in the latter during Roman times. The R1a lineage is believed to have originated in the Eurasian Steppes north of the Black & Caspian Seas. This lineage is relatively common in Scandinavia and in Slavic populations in Europe; those from Britain who are R1a are usually considered to have Viking ancestry.

According to Eupedia Haplogroup I is the oldest haplogroup in Europe. It is thought to have arrived from the Middle East as haplogroup IJ sometime between 45,000 and 30,000 years ago and developed into haplogroup I about 25,000 years ago. Nordtvedt estimates that haplogroup I1 split from Haplogroup I about 22,000 years ago but after 16,000 years only one male with haplogroup I1 survived to produce offspring. Thus, the most recent common ancestor (MRCA) for all of haplogroup I1 lived about 6,000 years ago. Haplogroup I1 is the most common I subclade in Northern Europe and is particularly common within Viking and Anglo Saxon populations. Nordtvedt has separated the haplogroup I1 individuals into a large number of subclades based on their STR DNA profiles and has calculated from the data the approximate dates when each subclade separated. Very recently the discovery of a number of new SNPs has helped to verify these divisions. Haplogroup I2 occurs less frequently in Europe than I1. The I2a haplogroup is common in the Balkans but less common in western Europe.

The foregoing discussion cannot be considered definitive concerning the deep ancestry of the participants but they do suggest what this ancestry might be. Most Hamiltons are thought to be derived from ancestors who originally lived in the area around Hamilton, Scotland, where both haplogroup I1 and haplogroup R1b1a2 families predominated.

Last updated by Gordon Hamilton June 2014