Modern census database searching is great. Many mis-indexed ancestors have been found by the ability to throw wild card variables into a tricky surname or, when all else fails, to abandon names altogether and search for characteristics like age, occupation and nativity.
But remember: Each page in a search result is just one possible piece of a family mosaic. Case in point:
I was scouring the 1870 index for the family of my great-grandmother Catherine Connors Haigney in Watervliet, Albany County, N.Y. By this point in my search I knew that Catherine’s oldest sister, Mary Ann, was likely to be married to a man named Bernard Connell in 1870. And there they were:
Excellent! (A bonus: They married in the census year, so the enumerator noted the month of their wedding, January. You can’t see it in this crop, but it’s there.)
Now it was time to check on my great-great-grandparents, Patrick and Bridget Connors. There was only one family in Watervliet in 1870 that included a head of household named Patrick, a wife Bridget and siblings whose names matched the known siblings of Catherine and Mary Ann. Up they popped:
Wonderful! There they all are, Andrew, Mary Ann, James … Wait.
Mary Ann? Seriously? But how could she be both the eldest daughter in Patrick Connor’s household and the wife of Bernard Connell? One finding had to be the wrong Mary Ann. Right?
I spent the next few minutes whimpering softly about what a rotten, horrible, deceptive world this is, where census indexes make us think we have a handle on a family, only to cruelly snatch our triumph away with the very next hit.
But soon I saw something that I should have noticed right away. See Bernard Connell and Mary Ann up there? See how they’re at the top of their page?
And see how Patrick and Bridget and their gang are at the bottom of their page?
Could these people just possibly be on adjacent pages?
You bet, Sherlock. The Connors and the Connells are, in fact, in the same dwelling, No. 727, but are enumerated as two distinct families, No. 902 and No. 903.
The Connors/Connell family group was visited by a somewhat persnickety enumerator in 1870, a year in which individual names were recorded, but relationships to head of household were not. Faced with the presence of Patrick’s married oldest daughter, the enumerator parsed the situation as precisely as he could. He listed Mary Ann first among Patrick’s children, and a second time as Bernard Connell’s wife. Then the entry happened to break across Pages 110-111.
There are not two 18-year-old Mary Anns in Dwelling 727. They are the same person whose dual identity has been carefully, if confusingly, preserved, a conclusion supported by other sources, including the obituary of one of Mary Ann’s daughters many years later. And, of course, these two Mary Anns appear as two separate census search results on separate pages, each seemingly valid, but contradictory. Only when the pages are read in sequence do they make sense.
It’s an elegant example of some basic census-research advice: Never simply zero in on one key name on a census page. Read up, read down and read adjacent pages. It’s the only way you’re sure you’re getting the whole picture.