An Unofficial Genealogy Rule: Always Check the Neighbors

This story began with a quiet family question. Nothing dramatic, just a sense that something in this branch of the family tree did not quite add up.

My father, an experienced genealogist, had spent years studying DNA results and inheritance patterns. When he reviewed this line in Ancestry ThruLines, one detail stopped him cold. The numbers were not behaving the way they normally should.

My grandmother’s grandfather was Luther Hobbs. The man Luther called his father and who raised him was David Hobbs. Luther’s mother was Adaline Hobbs. When my father looked at ThruLines, Luther had twelve DNA matches. David also had twelve DNA matches. Going back another generation, the number stayed exactly the same.

That is not proof of anything on its own, but it is unusual. Normally, ThruLines expand as you move backward in time. Each generation should introduce more descendant lines, more testers, and more matches. Here, the tree simply stopped growing. There are innocent explanations for this. Relatives may not have tested. Trees may be incomplete or private. Algorithms are imperfect. Still, even at first glance, it stood out.

Patterns like this are rarely proof on their own. But in genetic genealogy, they act like friction in the system. When something refuses to move the way it should, it is usually worth paying attention.

Years earlier, my grandmother had shared a story that might explain why. It was not framed as a scandal or a secret. Just something she believed might have happened, passed along carefully and without certainty. She believed that Adaline Hobbs may have had a relationship with another man before Luther was born. David Hobbs raised Luther as his own, but might not have been his biological father. Like many family stories from the 1800s, it was never written down and never proven.

For most families, that is where a story like this ends. Records are thin, witnesses are gone, and the social realities of the 1860s ensured discretion.

What made this case different was DNA. I was lucky enough to have my grandmother’s test results.

Always Start With the Oldest DNA You Have

Until recently, there was no way to test that story. Records alone were not enough. DNA changed that.

One of the most important principles in genetic genealogy is to start with the oldest DNA you have access to. Older DNA is closer to the question you are trying to answer and carries fewer generational splits. In this case, my grandmother’s DNA was the closest possible link to the people involved in the story. I did not begin with my own results or my father’s. I began by orienting myself with hers.

Step 1: Verifying the Known

Before looking for anything unexpected, I confirmed what should already be true. I linked my grandmother’s obvious DNA matches to documented relatives in her tree and verified that those relationships behaved as expected genetically. This step is not glamorous, but it matters. It establishes that the foundation is solid before you start pulling on loose threads.

Only once that confidence was in place did I move on.

Step 2: Narrowing the Dataset Deliberately

Because the question involved my grandmother’s paternal great-grandmother, I filtered aggressively. First, I limited matches to her paternal side only. Even after doing that, I was still looking at roughly 32,000 matches. At that scale, patterns disappear instead of emerging.

Next, I narrowed the list to second cousins using a custom shared DNA range of 30 to 200 cM. I wanted enough flexibility to capture a meaningful group without drowning in noise. That single decision reduced the dataset to about 300 matches. At that point, the problem became manageable instead of overwhelming.

Step 3: Clustering and Geography

With the dataset narrowed to paternal second cousins, I could finally focus on the unknown matches. My working assumption was that if Adaline Hobbs had a relationship with another man, he probably lived nearby, stayed in the area, and had descendants. Geography, in other words, was likely to matter.

Fortunately, Adaline’s location during the time in question was remarkably stable. She lived in one town in St. Clair County, Michigan, for several decades. So I searched my grandmother’s DNA matches for anyone with a birth in St. Clair County listed in their linked trees. The fifth and sixth matches on the list immediately stood out. That was where I started digging.

Step 4: Building the Cluster Forward

The first match I examined had a distinctive surname. Searching that surname across other matches’ trees quickly revealed three individuals whose trees overlapped. From there, I built a working tree for that DNA match and began linking him to everyone he was genetically connected to.

My logic was simple. If I could build a clean, well-documented tree for this cluster, eventually it would intersect with Adaline Hobbs in a meaningful way. I grew the tree three generations upward, adding only DNA-confirmed relatives and focusing exclusively on ancestors who lived in Michigan.

By the time the tree reached roughly 50 people, one man stood out immediately. He was the right age, he lived in the right place, and he fit the DNA evidence perfectly.

I pulled the 1860 census.

And there he was, listed exactly where the DNA suggested he should be.

The man who fit the DNA, the age, and the location perfectly was living next door.

The Conclusion

At that point, the pieces fell into place. The family story was true. My great-great-grandfather was not the biological son of the man who raised him, my great-great-great-grandfather. Instead, his biological father was the man living next door to his mother at the time of his conception.

The entire process took about ten hours. It worked because the scope was tightly controlled, the DNA was filtered intentionally, and geography was treated as evidence rather than background noise.

The practical takeaway is simple. When parentage questions refuse to resolve, start with the oldest DNA you have, control your scope carefully, and remember that geography is evidence. Sometimes the answer is not hidden in distant records. Sometimes it is right next door.