27

I have spent a considerable amount of time creating family trees on Ancestry.com, and documenting them with records found on Ancestry, FamilySearch, and in other web and non-web sources. For a variety of reasons, wound up with more than one tree for the same family (one tree focusing on my mother-in-law's family, and one on my father-in-law's).

In both trees I document not just the immediate descendants, but essentially anyone on whom I can get good evidence, so one of the trees is more like a little forest of inter-married families.

In retrospect, I am wondering if this compartmentalization is a good thing. Would it be better to have a single tree for everyone (containing over 1000 individuals), or separate trees for logical family groupings? What are some less-obvious tradeoffs?

If the answer is to have multiple trees, what is best practice for connecting them, making it easy to jump from one to the next?

How are these decisions different if (as is the case in my trees) there is inter-marriage between families at different generations, creating multiple paths between two family groups/trees?

Gene Golovchinsky
  • 8,071
  • 3
  • 30
  • 54
  • Ha! I have my paternal and maternal sides separated, too! – GeneJ Oct 12 '12 at 23:52
  • 1
    So how do you have your trees connected? And what happens if there is more than one point of contact, with distant cousins being married, or multiple intermarried families? – Gene Golovchinsky Oct 13 '12 at 00:10
  • 3
    What I do is I name my tree from the last names of all 4 grandparents, any in-laws get no slice of that cake but are still on the tree. – Mike Oct 13 '12 at 13:30
  • 1
    Mike's reference to "get no slice of that cake" made me smile. – GeneJ Oct 13 '12 at 20:05
  • My Sister made a family tree, but she started it before computers came out. The family trees became a very large amount of information. Then she got huge binder books to organize her work, when you do your research like she did, going back to the 1700's, you will find your family or relatives have a differen't last names She has it all, but she passed away in 2016 and I got all of her genealogy work. She put the families last name on each binder, and going through these huge binders, they are listed by generation. Cross family Campbell family and much more. – Patricia Luke Jul 17 '18 at 23:32

7 Answers7

15

I recommend that you keep all your own research in just one tree. Do not compartmentalize.

If any people in your tree are related to you more than one way, they will appear in more than one tree if you have more than one. Duplication will be a pain to keep synched.

The software should allow you to produce separate family reports relatively easily. It won't do as good a job if you're trying to use two or more trees to produce a report.

lkessler
  • 16,677
  • 2
  • 25
  • 81
13

I think this is another preference question with no correct answer. Though for me the advantages of having one big tree seem so clear and that's how I have always done it.

If it is important for you to be able to export GEDCOM files that are limited to one line of your ancestry then you probably prefer separate trees. Maybe there is some other reason that makes it important for you to compartmentalize your data.

I have a single tree on Ancestry.com, approaching 4,000 persons, that includes all the people I have researched via Ancestry. Some of them are disconnected from the others because I haven't yet found enough information to link them into known lines. I have not found any disadvantages of this approach.

In my LifeLines database, which is decades older, I have about 16,000 person records. A good number of these are what are sometimes called "persona" records, so the database probably has about 14,000 "conclusion persons." Again I have only one single database for all my data.

If I were a professionaly genealogist, I would have a tree for each client. And when I do work for friends I keep their trees separate. But for my own tree, which is also combined with my wife's ancestry, as well as the ancestry of my sons-in-law and daughter-in-law, since they are ancestors of my grandchildren, it all goes into one big tree.

In addition, my single tree is also growing into a one-name database for the Wetmore surname. There are no disadvantages that I have ever felt for keeping all this data in a single "tree." I put tree in quotes here because obviously this isn't a simple tree. I basically go whereever the data goes.

On the Ancestry.com side, when you export GEDCOM, I'm not sure if you can limit the sections of your tree that are exported. However, with LifeLines I can fine tune exactly what part of the database I want exported. This way it is trivial to create and send a custom GEDCOM file with a precise subset of the overall database. If I want to send a subset of the Ancestry.com database, it is easy to export the whole thing as one GEDCOM file, import that GEDCOM file into a fresh LifeLines database, and then use LifeLines to "re-export" the precise part of the entire tree that I need.

Tom Wetmore
  • 3,162
  • 20
  • 24
  • When I started the second tree, it was initially a way to document a bunch of people with the last name Decker who lived in Chicago and had come from a particular town in Germany. I figured they might be related, but didn't know how. Since then, I have figured out how they are all related, and now I suppose it makes sense to merge them with the other tree that contained only known relatives. – Gene Golovchinsky Oct 13 '12 at 03:07
7

Many different approaches and recommendations here. It's a subject with no definitive answer.

It's worth noting that "tree" is, in general, an inappropriate description for family lineage because real-life is never quite as linear as we'd like. The technical term would be a Directed Acyclic Graph (DAG) since - as far as I know - no one has found a way of being their own ancestor (i.e. no loops in the lineage). All other types of cross linkage are possible though.

As someone who follows all their family lines (i.e. both pairs of surnames in each generation), I know the dilemma you're facing. I compartmentalise my data in offline storage which helps with sharing but doesn't impact maintenance. References to a person in a different dataset (e.g. a spouse whose family has their own dataset) are represented by special cross-dataset links. However, I write my own software to support this model so I cannot suggest it as a solution.

The point I'm trying to make here is that the approach depends on where you store your definitive data. Do you use unindexed offline storage such as GEDCOM files or written documents? A lot of people probably pour everything into some database with a proprietary schema but I believe that it is risky and constraining to treat that as definitive.

ACProctor
  • 5,897
  • 20
  • 37
  • DAG (a Directed Acyclic Graph) sounds like something I should remember (but don't) from long-ago Pure Mathematics. Does this model assume that nodes (or vertices) are unique? I am sure that I am not the only person with a marriage of second cousins in their family, so that (one set of) the grandparents of that couple appear twice (occupy two nodes). – Fortiter Oct 14 '12 at 12:10
  • What did you mean by unique nodes? If every person constitutes a node in the DAG then I'm not quite sure what you mean. Unrelated note: it's the temporal nature of generation that makes it a DAG. Short of time-travel, a loop cannot happen ;-) – ACProctor Oct 14 '12 at 14:11
  • The assumption of unique nodes in a classic family tree is that each person appears at one and only one vertex of the graph. Consider this example of Prince William's tree http://en.wikipedia.org/wiki/Ahnentafel#Example in which "persons" 80 to 83 are already in the tree at other locations (nodes). I wonder if graphing that situation breaks the assumptions of a DAG? – Fortiter Oct 14 '12 at 14:26
  • Tony is right. Every "conclusion person" occupies one node in the "tree", which will become a directed acyclic graph when the first cousin marriage enters the picture. The fact that the same person shows up in multiple places in a pedigree chart with cousin marriage has nothing to do with whether they occupy two nodes in the underlying graph -- they do not. – Tom Wetmore Oct 14 '12 at 18:02
5

The biggest risk of keeping separate trees is that an individual may (unwittingly?) appear in multiple branches, resulting in multiple sets of incomplete data, or even worse multiple sets of inconsistent data.

It is bad enough trying to correlate individuals when they all appear in a single list...

Andrew
  • 726
  • 6
  • 19
4

Although we all use the term family tree, I suspect that very few family historians actually maintain pure trees in the genealogical sense. Software tools have made it so easy to expand the number of individuals about whom we keep records far beyond the set of direct ancestors.

Long ago when I was teaching introductory ecology to secondary students, I saw them struggle with the shift from a neat linear food chain (A is eaten by B which is eaten by C) to the real-world complexity of a food web (where D, E and F compete to prey on both B and C). Most of the data collections that we think of as trees are actually richly inter-connected family webs.

The "forest of trees" image is actually misleading because when you try to imagine "your" tree side by side with those of your third cousins there will be a dispute over whose tree 2xgreat grandmother belongs in. Can one limb be growing on two trees?

Once you think of your collections in these terms, the question becomes do I want to keep multiple snapshots of different parts of this web or to consolidate them into one representation that captures all the (known) complexity?

And then, you must set philosophic considerations aside and determine if joining the datasets is actually practicable in your case! You don't want to destroy a workable system in pursuit of an ideal.

Fortiter
  • 13,336
  • 1
  • 20
  • 69
4

I find one tree simplest to maintain, even though I am (or may be) researching unrelated individuals. I find it:

  • eliminates the risk of recording different information about the same individual in different trees
  • reduces the risk of inadvertently recording the same individual twice
  • avoids having to copy data from one tree to another if I discover the same individual 'belongs' in more than one tree
  • makes it easier to maintain a consistent list of sources, as I only have to record each source once
  • makes it easier for me to know that I'm working in the 'right' tree!

I once maintained separate trees for "individuals I knew were related to me" and a bunch of "other James families from the same village as my lot who might or might not be related". When I discovered that some of the other James families were in fact related, it took great care and attention to merge some but not all of the second tree into the first, so in the end I merged everything and haven't regretted it.

However, if Ancestry (or an equivalent software package) have limitations on tree size, that would be a practical consideration to take into account, as would the ability to report on or export just part of the tree when sharing informatin with other people.

2

My personal opinion is that a pedigree or family tree should be kept as a whole; however, let's look at the possibility that there is a branch of the family that you can't link through documentation. This is where I would chose to make a separate tree. I have observed paternal trees, maternal trees, a grooms and/or bride's line. These are a personal preference, if someone chooses to record their trees in this way it's neither right nor wrong.

Ezri Rediker
  • 2,232
  • 12
  • 15