Thursday 29 May 2008

a1b4c5d2e8f7. WHAT DOES IT MEAN?

Colin Möhr recetly asked the above question on Rootsweb (South Africa). A long discussion followed. Here are some edited replies.

A while ago it was announced that approximately 800 000 South Africans emigrated over the last decade. Johann Hanekom indirectly refers to such by writing: ... and so they are becoming progenitors in their own right in their new abodes. Does it make genealogical sense to hang on to their SA coding?

Although lengthy, this articles is a must to read --
for beginners and experienced genealogists.

Richard Ball, Norfolk, England: First of all, always bear in mind that the De Villiers numbering system, which you are quoting here and which is used in the original Villiers Geslagtregister', also De Villiers/Pama revision of 1966 and the current volumes of the the South African Genealogies, still in course of publication, is a relative system, not an absolute one.

Therefore, any descendant number such as you quote above must first have

  • a family name and then
  • publication to which it refers.
It is not valid in the abstract,even though two publications may agree on the generations, you should always also quote the publication from which you are quoting the generation code.

For more information on this coding system see: The numbering system we are discussing is on that page 6 -- the de Villiers/Pama system but in a nutshell, here is how it works:

a is the progenitor or stamvader, the person who first arrived in South Africa. There may have been just one, in which case the a is not usually quoted (although the surname must always be quoted) or there may have been more than one progenitor, in which case they will be labelled a1, a2, etc.

b is the second generation (or first one born in SA -- sometimes some confusion here) and the number is the order of that particular child in the birth order. b6 is the sixth child of the progenitor.

b2c3 is the third child of the second child of the progenitor, and so on.

These numbers do not carry on to the children of female descendants which, in the De Villiers system, take their numbering from their father's family.

To illustrate what I mean when I say the code is only relevant to the publication in which it appears, take the Van Wyk family where there were 6 progenitors, including the two main early ones, where the 'Die VAN WYK-families van Suid-Afrika', compiled by the Noord-Transvaal branch of the GSSA, labels Willem van Wyk, a2 and Roelof van Wyk a1.

This is a recent revision of the family and differs from what appears in the original Geslagtregister' or 'DeVilliers/Pama' and any numbering of the generations will only be intelligible if the publication from which the code is taken is also quoted.

For instance take the following two Van Wyks:

the Roelof van Wyk who married 1727 Aletta Bezuidenhout is coded a2b1 in the DeVilliers/Pama book, but in the new revision a1b1c1;

the Willem van Wyk who married (1)1729 Johanna Campher and (2)1746 Hendrina Monk (and left many descendants) is coded a2b3 in the DeVilliers/Pama book but in the new revision a2b6.

So, unless you also have the name of the publication, you cannot identify an individual simply with a code.

Francois Greeff: Think about the kind of "roadmap" you get when you stop to ask for directions in a city:

"Take the second left
then the third right
then the second right
then the fourth left
then the seventh turning to the left again..." and so on.

Effectively, you have a set of directions that get you from one specific point in the city to another specific point.

A family tree is just a map of how the different points in a family are related to each other, and there are two different kinds of map: One for ancestors, and another for descendants. The reason for this is that the two trees, or maps, have different structures.

A map of ancestors follows a very rigid progression because each person has exactly TWO biological parents. Each ancestral map has the same structure (from the bottom upward):

128 ggggg grandparents, and so on (each generation having twice as many people as the one before)
64 gggg grandparents
32 ggg grandparents
16 great great grandparents
8 great grandparents
4 grandparents
2 parents
1 person

This structure is easily seen in an ancestral chart (Click on the link to see an example, and count the number of people in each column:

In this tree each person has a unique place AND NUMBER that defines his or her relationship to all other people in the tree. The numbers are simply: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 11, 12, 13... and so on. (See

If you click on the last link you will see that it happens that:

1. all males have even numbers (except number one, who could be female);
2. all females have odd numbers;
3. each person's father is double their own number;
4. each person's mother is double their number, plus one and
5. every wife has her husband's number plus one.

The structure of this ancestral chart (or map) is EXACTLY THE SAME FOR ALL PEOPLE. One can thus refer to the chart and say that numbers 1, 3, 8, and 9 attended the funeral of number 19 and any person in the world could see that the person whose tree it is (1) and his mother (3) and his paternal grandfather's parents (8 & 9) went to the funeral of number one's father's father's mother's mother. Using the numbers to refer to people is clearly much simpler than "number one's father's father's mother's mother", especially when one gets to a huge tree of ten or twenty generations.

Second and third wives and husbands NEVER enter into this kind of ANCESTRAL TREE for the simple reason that the tree lists ONLY the ancestors of number one, and no one can have two or more biological mothers.

Now we can look at a map of descendants.

Descendant's trees (or maps of family relationships) are far more difficult, simply because each set of parents has a different number of children, and different marriages.

In this system we have two separate counters:

  • One counter for each generation (a, b, c, d, e, f, g, h, i, j and so on) and
  • Another counter for the number of children in THAT family (1, 2, 3, 4, 5, etc).

In this system ONLY BLOOD RELATIVES ARE GIVEN NUMBERS, and by that I mean blood relatives of the one person whose descendants we are describing.

The DESCENDANT'S CHART also starts with ONE PERSON, who is coded a, because a is the first generation in the tree. It is a common mistake to put the progenitor (or starting person in the tree) down as a1, because that means that he was his parents' first, or oldest, child. If he is known to be the fourth child, then he should be a4.

So a has two wives, who have NO NUMBER in this system. a has four children by his first wife, and three by his second wife. All of a's children are the second generation in the tree, so they are generation b, and in addition to that they are numbered in order of birth: b1, b2, b3, b4, b5, b6, b7.

It is obvious that b5 is the fifth child of a. Now b5 has five children, who are generation c in this tree. They are c1, c2, c3, c4, c5.

A problem arises immediately, because b7 also has children, six of them, and in this tree they are also the third generation, thus c. They are ALSO c1, c2, c3, c4, c5, c6.

In order to tell all the c generation children apart we look at who their parents and grandparents were, and renumber them:

ab5c1, ab5c2, ab5c3, ab5c4, ab5c5 and
ab7c1, ab7c2, ab7c3, ab7c4, ab7c5, ab7c6.

Now each person has a unique number.

That unique number also maps out each person's PLACE IN THE TREE (or place on this family map).

Returning to the example you asked about in the first place, "a1b4c5d2e8f7", you can begin to see that this long number is:

  • A specific person in a tree and
  • A set of directions that define his relationship to all the other people in that tree.

Do you remember the set of directions above you got to travel across the city? "Take the second left, then the third right then the seventh turning to the left again..." and so on.

"a1b4c5d2e8f7" is also such a set of directions for you to navigate in a family tree (or a map of the family). In fact, an abbreviation of a set of directions. Let us break "a1b4c5d2e8f7" up and look at what "a1b4c5d2e8f7" means.

  • a1 is the main person around whom the entire tree is built, a because he is the first generation, and 1 because he is the first child of his parents.
  • b4 is the fourth child of a1
  • c5 is the fifth child of b4
  • d2 is the second child of c5
  • e8 is the eighth child of d2 and
  • f7 is the seventh child of e8.

This system was invented by the late de Villiers to describe SOUTH AFRICAN FAMILIES. De Villiers, consequently, made an exception of person a, because he was studying a very specific group of first generation people. All of his first generation people were THE FIRST MAN OF A FAMILY TO ARRIVE IN SOUTH AFRICA -- the progenitor of his clan in South Africa.

Very soon de Villiers ran into a new problem, of which my clan happens to be a good example:

  • Matthias Greeff arrived at the Cape in 1680;
  • Peter Greeff arrived at the Cape in 1734 and
  • Friedrich Greeff arrived at the Cape around 1775.

So now we have THREE progenitors of "the Greeff family" in South Africa.

De Villiers solved the problem by making an exception of the number that follows:

a: Given that he had three Greeff progenitors he numbered them a1, a2, and a3.

Consequently the numbers behind a DO NOT mean that that person was the first, second or third child of his parents, but it means that he was the first, second, or third person of that surname to arrive in South Africa.

In the De Villiers numbering system the number following a is thus NOT a mistake, because it still refers to the first, second or third person (or progenitor - "stamvader") to arrive in South Africa.

For the sake of simplicity I have not covered some of the other exceptions in the system, nor all the finer details.

Johan Hanekom: It is a valid point about always quoting the publication from which the code for an individual is taken -- but lately it has became apparent that the REVISION of that publication also needs to be considered, as revisions correct errors and omissions from previous revisions. If I do my job on our family tree well, I am afraid that the next revision of SAG Vol3 is going to be very interesting to see.

My interest in the Hanekom family tree ( goes way beyond the South African progenitor, and I am sure that goes for a lot of researchers here too. Unfortunately that makes a bit of a nonsense of the coding.

Richard Ball: Oh yes, certainly -- not only the publication but its edition, too. The coding is only meaningful in the context of its exact publication.

Johan said: “My interest in the Hanekom family tree ( goes way beyond the South African progenitor, and I am sure that goes for a lot of researchers here too”.

I presume you mean you are interested in the the progenitor's ancestry? This is seemingly a modern concept. It seems the earlier researchers had little interest in the European or other antecedents of the stamvaders.

Johan also said: “Unfortunately that makes a bit of a nonsense of the coding”.

The De Villiers coding seems firmly established for the main South African genealogical publications. I have seen a variation that goes backward from the progenitor using upper case letters B1.C2 etc, but you are, of course, at liberty to use your own for any publication you produce.

Johan Hanekom: I think I am not unique in finding that more and more originally South African family names are no longer purely South African family names, but becoming international. In my research into the Hanekom family I find a lot of members now living in the UK, USA, Saudi Arabia, Namibia, New Zealand, Australia, and even one (my son) back in Germany, close to where our SA progenitor came from originally. And so they are becoming progenitors in their own right in their new abodes. Does it make genealogical sense to hang on to their SA coding? I don't know.

With the excellent charting and reports now available I personally find it easier to use a pedigree chart to display and track the ancestry of a given person, than trying to implement a numbering system that keeps changing.

Now let me throw a second stone in the water. At the risk of offending researchers who publish their family histories, I would posit that genealogical books don't generate a lot of royalties for the author, and that e-books containing the anecdotal history of the family, with a searchable family tree, would be a worthwhile consideration. It would be cheaper for the author to publish, easier to update, and of much more value to genealogical researchers -and save a few more trees in the process.

Richard Ball: Villiers, however, if indeed he devised the system (his incomplete work was edited and prepared for publication by G.M.Theal), was not designing a database but a system which would help a reader to locate the various generations in a long printed list, similar to such systems as used by most genealoly programs. If you look at the original books that G.M.Theal produced from De Villiers's data, you will see the logic for

1. the printed page
2. 5 or 6 generations
and part of this original system was the use of specific indentation for each generation, as below, making it relatively easy to run one's eye down the page, concentrating on one generation. (Dots inserted for the sake of neat typing.)

This was the purpose of the system.

No comments:

Post a Comment

Note: only a member of this blog may post a comment.

Related Posts Plugin for WordPress, Blogger...