Everyone has one. The ancestor who appears in exactly one record and then stops existing. The great-great-grandmother with no maiden name. The line that runs beautifully for four generations and then hits 1847 and dies.

Most advice about brick walls is a list of tips, which is useless, because the tips don't apply equally and you don't know which one you need. What's actually useful is figuring out which kind of wall you've hit, because there are only about five, and each has a different way through.

The five walls

1. The name changed

The most common wall, and the one people least expect, because we treat names as fixed and our ancestors didn't.

Spelling was fluid. Clerks wrote what they heard, and an illiterate family had no way to correct them. One family can appear as Reilly, O'Reilly, Riley, Rielly, and Ryly across five documents, all correct. Emigrants anglicized names deliberately — Schmidt to Smith, Kowalczyk to Kowal or Kovacs, Giovanni to John. Names were translated rather than transliterated, so a German Johann becomes John and a Welsh Ieuan becomes Evan.

And people used middle names. A man baptized John Thomas Whitfield may appear as Thomas Whitfield in every census of his adult life, because that's what everyone called him.

The way through: search by everything except the surname. Search the given name plus a birth year range plus a place. Search the spouse. Search a child with an unusual first name. Use wildcards aggressively where the search supports them. And say the name out loud in the accent of the person who would have been speaking it — a surprising number of variant spellings become obvious that way.

2. The record exists but isn't indexed

This wall is invisible, which is what makes it dangerous. You search, you get nothing, and you conclude the record doesn't exist. What actually happened is that the record was never transcribed into a searchable index, so no amount of searching will surface it.

Enormous quantities of digitized material are in this state — images sitting online, unindexed, findable only by browsing them page by page.

The way through: stop searching and start browsing. Find the specific collection covering the right place and time, and go through it. This is slow and it is how a great deal of real genealogy gets done. FamilySearch's catalog is the best tool for identifying what exists for a given locality, including material that is digitized but not searchable, and material that is only on microfilm. Most people never open the catalog and it is the most powerful thing on the site.

3. The record was destroyed

Sometimes the answer is that the document is gone, and knowing that saves you years of looking.

The big losses worth knowing:

  • The 1890 US census. Damaged by fire in 1921 and disposed of in the 1930s. A twenty-year gap between 1880 and 1900 that swallows an entire generation of American research.
  • The Public Record Office of Ireland, 1922. Destroyed during the Civil War at the Four Courts, taking centuries of Irish records with it, including most pre-1901 census returns. Irish research is shaped by this event more than by anything else.
  • The National Personnel Records Center fire, St. Louis, 1973. Destroyed a huge share of 20th-century US Army and Air Force service records.
  • Courthouse fires. Endemic across the American South and Midwest in the 19th century. County-level records — deeds, wills, marriages, probate — frequently just end at a particular year.

The way through: substitutes. Almost nothing is recorded only once. State censuses, city directories, tax lists, voter rolls, church registers, land and probate records, newspaper notices, fraternal and union records, school censuses. A destroyed census is not a destroyed life; it's one missing document among many that were generated by the same person.

4. The place changed

You're searching the right name in the wrong archive, because the jurisdiction the record belongs to isn't the one the place sits in now.

Counties split. Towns were absorbed. Borders moved wholesale — a village in what is now Poland generated its records in German, in an archive that may now be in a third country. And the "hometown" written on an immigration record is frequently not a town at all but a parish, a province, or whatever a clerk heard.

The way through: research the place before the person. Find out what jurisdiction held the records in the year you care about, and work the ladder — parish, then civil registration district, then county or province, then state. If the place name on your document doesn't appear on a modern map, that's a specific and solvable problem rather than a dead end.

5. The assumption is wrong

The hardest wall, because you can't see it. You're searching for a person who didn't exist in the form you're searching for.

The usual culprits: a birth year that's off by five or ten years, because ages in censuses are approximations and people lied about them routinely. A father who was actually a stepfather or a grandfather. A "first" marriage that was a second. A family that never lived where everyone says they lived. A relationship copied from someone else's tree in 1998 that was a guess and has since been repeated into apparent fact.

The way through: go back to the last thing you actually proved and audit everything after it. For every link in the chain, ask what document says so. If the answer is "another tree," that link is unproven — and there's a decent chance your wall is built on it rather than in front of it.

The technique that actually breaks walls

Here's the thing most articles bury, and it's worth more than the five categories above.

Stop researching your ancestor. Research everyone around them.

People did not move through history alone. They migrated with siblings and cousins and neighbors from the same village. They witnessed each other's weddings. They stood as godparents. They were named as executors and bondsmen. They bought land from each other and appeared on the same page of the same census, ten years running, in two different states.

This is called cluster research, or the FAN principle — friends, associates, neighbors — and it's the difference between amateur and competent genealogy.

The mechanism is simple: your ancestor is invisible, but the people around them aren't. A brother who stayed put and left a well-documented paper trail can name a birthplace your ancestor never recorded anywhere. A marriage witness turns out to be a first cousin whose own records identify the shared grandparents. A neighbor in Ohio turns out to be a neighbor in Bavaria twenty years earlier, which tells you exactly which parish to search.

In practice: take one document your ancestor appears in and research every other name on it. Witnesses, informants, neighbors on the census page, the other family in the same household, the person who posted the bond. It feels like a detour. It is the most reliable technique in the field.

Habits that prevent walls

Record where everything came from. Not just what you found — where. A conclusion without a source is a conclusion you'll have to prove again in five years, and it's how bad information spreads between trees.

Distinguish what you've proven from what you've assumed. Both belong in your research program; they should not look identical when you come back to them.

Don't copy from other trees without checking the sources behind them. An unsourced relationship in someone else's tree is a hypothesis. Copying it makes it a hypothesis in two trees.

Write down what you searched and didn't find. Negative results are real findings, and without a record of them you will search the same collection three times over four years.

When a wall finally falls

It usually isn't a moment of insight. It's a name in a witness column that means nothing until, six weeks later, it means everything.

And the payoff is real. A line that stopped at 1847 suddenly runs another three generations, and a section of the chart that was blank for years fills in — which, when the whole thing is finally on a wall or a screen where the family can see it, is the part people actually respond to.

The wall you're stuck on right now is almost certainly one of the five. Work out which, and there's a way through it.

Keep reading: