I used to collect people. Names, dates, places. It looked like progress because the boxes filled. Then a crosswind hit and half the branches bent the wrong way.
This morning I sat down at 7:10 with the 1910 census again. Same woman, same wrong age. My great-grandmother shaved off two years in that count and kept the story. I’ve excused it before. Not today. Today was about resolving conflicting ages across census, headstones, and parish records without kidding myself.
Notebook out. Three columns: Source. What it says. Why I believe it or don’t. I pulled only what I actually have, not what I remember having: a parish marriage entry (exact parish withheld until I confirm with the register images), two city directories, the 1900 and 1910 census, a burial-permit index, a headstone photo with pretty numbers that do not earn trust.
The directories were the pivot. People treat them like dusty phone books. They are better than that. In 1908 she is “Mrs. M. J.” on Pine Street, occupation “seam.” In 1909 the house number shifts and the boarders change. That fixes her in time and space more tightly than a fuzzy census age. Using city directories as genealogical proof sounds grand; in practice it is a clerk with tired shoes walking a route and writing what is painted on doors. Often that is the least wrong line in the whole pile.
By 8:30 the light had crept across the desk. Knees complained. I deleted the fake birth year and replaced it with a window bounded by the marriage entry and a likely baptism span based on local practice. Then I wrote each claim as a plain sentence a decent skeptic could read without rolling their eyes. “In April 1910 an enumerator recorded Mary as 22.” “In June 1907 a parish register recorded Mary’s marriage, age 21, father Patrick.” “In 1908 a directory lists her as married and resident on Pine Street.” Then I ranked them by why they would be right or wrong. Ceremony in front of a priest beats a distracted headstone cutter. Route-walked directory beats a memory thirty years later. Simple. Not easy.
Madrid footnote, because life leaks in. The pharmacist downstairs raised an eyebrow when I bought more finger splints. She has started calling me “Profesor.” I am trying not to enjoy it.
I checked the neighbors in 1910. Two doors down: a widow from County Mayo with a brother I had seen as a marriage witness. Same rare middle initial as my great-grandmother, which I have always side-eyed. In 1900 that brother is in the widow’s household. The witness ties to the neighbor. Cluster research, not magic. People move in clumps.
At 9:12 the phone rang. An old colleague from Los Angeles. We did a brutal week together in 1994 on a courthouse microfilm reader; I still hear the thunk of that machine. He sounded rattled. Bad car accident on the freeway. Airbags, glass everywhere. “For a second,” he said, “I thought I was about to become an index entry.” Dark joke. I got it. We talked through shock, insurance, and documented timelines, because paper helps even when it is not genealogy. I told him to speak to a proper Los Angeles car accident lawyer and to write down everything while the details were still sharp. Then I made myself a note: when you research strangers for a living, keep your own records straight.
Back to the desk. I opened a research log. Date. Objective. Search. Result. Next step. Objective: resolve the age conflict and test the neighbor–witness link. Searches: city directories 1907 to 1911; parish marriages 1907 to 1909; census 1900 and 1910; burial permits 1915 to 1930; draft cards for the brother to lock addresses. Results: enough to stop the tree from lying. Not enough to strut.
I am not naming the parish here on purpose. I do not cite what I have not verified with my own eyes. When I have the images, I will name the church, volume, and page. Until then it is a working label, not a fact. That is the part most trees skip.
Field notes, today:
- I trust the marriage entry. Contemporary record, created at the time. It names the couple and her father. I will add church, volume, and page once I pull the images.
- The 1908 to 1911 city directories fix her on Pine Street and show the job shift from “seam” to “dressm.” They line up with later census addresses.
- The 1910 census age and the headstone year do not hold up. Both get low weight.
- The marriage witness shows up two doors down in 1900 inside the widow’s household. His draft card later lands on the same street. That ties the cluster.
- No baptism in the indexed range. Next step is the unindexed books for 1885 to 1888.
How I am working it: write every claim in plain English, rank each record by why it could be right or wrong, follow the neighbors and witnesses, and keep a tight log so I do not drift.
What I can stand behind today: the headstone year is wrong; the census age is noise; the marriage entry and directories, read together, give a defensible birth window. The neighbor’s household and the marriage witness are the path to the maternal line. The plan is dull and perfectly fine. Page through unindexed baptisms for 1885 to 1888, map Pine Street households 1900 to 1912, order the widow’s probate packet.
I shut the notebook at noon. The van outside stopped beeping. Knees still sulked. The tree is smaller and stronger. There is a probate file somewhere with my answer sitting quietly in it. That will do for today.