The Elusive Sarah Gilbert

I've blogged quite a lot about Sarah Gilbert Johnson, wife of Ezekiel Johnson, and my great-great-grandmother. She appears in the marriage records of Clay County, Missouri, at the time of her marriage to Grandpa Zeke and in the 1880 census for Kansas City, Missouri. Then she appears of record no more.

I've been over and over the available materials. Indeed, a key purpose of my big research trip to Missouri this past summer was to find more on Sarah Gilbert. I found nothing new. I talked to several cousins descended from Zeke and Sarah Johnson, but none knew anything about Sarah.

Here's what I know about her: she and Zeke Johnson were married in Clay County, Missouri, on September 5, 1867. On the 1880 census, her age is given as 31. If that's true, she would have been born in about 1849. Her children are listed as Henry, 11; Mary, 10; Richard, 8; Ambrose, 6; Robert, 4; and Mattie, 1. I've since discovered that Mary (my great-grandmother) was in fact the oldest and not Henry. "Ambrose" was actually named Amos. I have death certificates for Mary, Richard, Amos, and Robert. Of Henry and Mattie, I know nothing beyond the 1880 census data. [Actually, the problem with Henry is that "Henry Johnson" is such a common name. I have identified some records that are "possibly" the "right" Henry Johnson].

Grandpa Zeke died in 1933. But in 1920, he was living with Richard and listed as widowed. So Sarah must have died before 1920. Noting that there seems to be no death certificate for her in Missouri's excellent 1910-1956 death certificate database, one might surmise that she died before 1910. This is a situation in which an 1890 census would be a great help. Her last child seems to be Mattie, born in 1879. She would have been just thirty years old. Perhaps then she died between 1880 and 1920.

There is one other strange thing that appears concerning Sarah Gilbert Johnson. There is a death certificate for one Robert Franklin Johnson who died on May 17, 1955. One might presume that this would be Sarah and Zeke's son, Robert, who according to the 1880 census, was born in about 1876. But the birth date on the death certificate is November 23, 1891. And the mother's maiden name is given as "Sarah Agnes Lewis."

The place of death on the Robert Franklin Johnson death certificate is given as "2444 Chestnut Avenue" in Kansas City, Missouri. On May 17, 1955, that house would have been occupied by my great aunt Rosetta Bell Long, and perhaps my grandmother, Annie Florida Corrine Long and her daughter, my aunt Delorise Gines. Rosie and Flo Long were sisters and the granddaughters of Zeke and Sarah. In May, 1955, Grandma Flo would have been 53 years old, and Aunt Rosie would have been 55 years old. And indeed, the informant is stated to be "Rose Long" of 2444 Chestnut Avenue.

But why did Aunt Rosie get Robert's birth date wrong by 15 years? And who is "Sarah Agnes Lewis"?

Yesterday, Lorine McGinnis Schulze blogged about the potential inaccuracies of primary records. She gives the (actual) example of a 17 year old informant who may have been confused by the questions asked by the official making out a death certificate. But Aunt Rosie was a mature, educated and experienced woman. How could she have made these mistakes?

The week before Robert Franklin Johnson died, my grandfather, William Edward Gines died. Could Aunt Rosie confused some of his information with that of her uncle? The informant on my grandfather's death certificate was "Harry Gines" [most likely his brother Henry]. Harry got everything right. My grandfather was born on August 10, 1898, and his mother was Sylvia LeJay [which is misspelled "LaJay" on the death certificate]. None of this information is similar to Robert's.

So what happened here? How did Sarah Gilbert manage to elude me once more?