Friday, August 18, 2017

Skepticism and the Census

Anyone who has spent much time chasing their ancestors in the census knows that you have to take the "facts" on a census return with a few fistfuls of salt. A lot of our forebears, especially the ladies, seemed to view "age" as more of an abstract concept than an objective fact, either not knowing, not caring, or actively lying about their age (how many of yours, remarkably, aged less than 10 years between consecutive censuses?).1

Then there's the reality that, except for the 1940 US census, we have no idea who actually answered the enumerator's questions – perhaps a step-parent, a 12-year-old, or a handy neighbor. Good luck with getting valid information there.

And of course, there are the enumerators themselves: they sometimes missed individuals2 or whole families,3 applied great creativity to the spelling of names,4 often had atrocious handwriting, and no doubt compounded these problems with copying errors.

But did you ever consider whether the pages were even numbered in the right order? I'm not talking about pages being microfilmed or digitized out of order, or missing pages – that's a whole separate issue. And I don't mean the often multiple numbers confusingly stamped or scrawled on pages after whole cities or counties were assembled. No, I mean when the pages are actually numbered 1, 2, 3, and so on, within a single enumeration sub-district or similar entity, and appear in that order on the digitized microfilm. I always pretty much assumed that could be relied on.

Until the other day, while I was working on my Saint-Cyr line in Quebec's Eastern Townships.

For those who aren't familiar with the 1861 census of Canada East (Quebec), it's a fully nominal census (i.e., all household members are named, not just heads of household), enumerated on 50-line sheets. However, there are no dwelling or family numbers, and no relationships to the (unidentified) heads of households. You have to deduce where a dwelling begins from the columns describing what the house is built from and how many stories it has, and where a family/household begins by how many families live in the house and who has an occupation listed. To further complicate matters, the enumerators were inconsistent (at least where my folks lived) about the order in which they listed family members after the apparent head and apparent spouse: sometimes they're in decreasing age order; sometimes boys first, followed by the girls, or vice versa; if a household member was absent at the time of the census (there are columns to indicate that), they might be listed at the end of the household, or mixed in with the others in whatever order they were in.

With that background out of the way, there were two "Sinsire" households (a father and an adult son) in district 2 of Arthabaska Township, Arthabaska County,5 the father's household (all surnamed Sinsire) occupying lines 45-50 at the bottom of page 3 and the son's (mixed Sinsire and Bélanger, from the wife's previous marriage), lines 44-50 at the bottom of page 7. Of course I checked page 4 to see if there were any more "Sinsires" at the top – there weren't – and page 7 was the last page of sub-district 2, so there were no more in the son's family. I thought.

Then I moved on to the 1871 census, where I discovered the son had two additional children, both well over 10, who should have been there in 1861, but apparently weren't.6 Was there a missing page 8 in Ancestry's database? (There was no page 8 or 9, with district 3 starting at page 10.) I checked on Library and Archives Canada. There was no sign of a page 8 there either. I checked that specific district for "Sinsire" entries, and, oddly, turned up two additional names, neither of which were the missing ones from the son's household. They were aged 14 and 11, listed on lines 1 and 2 at the top of page 6... but the last household on page 5 started with the surname Croteau on line 50. I began to smell a rat.

One of the missing children from the son's family was named Felix, so I searched for the first name in that district. Lo and behold, there was a Felix Bélanger – I had been looking for a Felix Sinsire, because in 1871 he was listed with the surname St Cyre – as was everyone in the household including the known Bélangers. Along with Felix Bélanger in 1861 was Edèse Bélanger. My other "missing" child found in 1871 was Thedèse St Cyre. And both of these kids were exactly the right ages. That was when I realized that they were too old to have been St Cyrs: they had been born before their mother had been widowed and remarried. Like the two mysterious extraneous "Sinsire" children, the two Bélangers were at the top of a page, in this case page 3, while page 2 ended with a Bilodeau family.

In short, the seven pages for the district had been numbered 1 to 7, but in apparently random order. By examining the names at the tops and bottoms of each page, I determined that the proper order of the pages is as follows:
  • 1 – Laliberte ... Barbier7
  • 7 – Turgeon ... Sinsire [fils]/Bélanger
  • 3 – Bélanger ... Sinsire [père]
  • 6 – Sinsire ... Frenette
  • 4 – Frenette ... Perrault
  • 2 – Perrault ... Bilodeau
  • 5 – Bilodeau ... Valliere/Croteau (2 families, 1 house)
Once I had the pages in the right order, all became clear. The two Bélanger children were the "St Cyre" children in the son's family in 1871, and the two extra "Sinsire" children belonged to the father's family.8

I have no idea if the rest of the districts – or the rest of county or province, for that matter – are as confused as this one small chunk. But I'm sure going to keep this in mind for the future: checking the "next page" for additional family members may give you a false negative – the page numbers may simply be wrong, and those additional members may be on an unexpected page.

  1. I have at least one who miraculously managed not to age at all between two censuses. 
  2. My father and his sister, aged 8 months and just over two years, respectively, were apparently not in residence with their parents when the census taker came around in 1920. Where else they would have been in January, in Maine, I can't imagine. I also can't imagine my grandmother just forgetting to mention them.
  3. Like, say, my mother's entire family in 1930.
  4. I can see maybe having trouble with an uncommon name like "Sukeforth," but you would think it would be hard to foul up "Kirk." You would be wrong.
  5. 1861 census of Canada East, Arthabaska County, Quebec, district 2, Arthabaska township, pages 1-7; database and digital images, Ancestry ( : accessed 9 Aug 2017).
  6. I also had some issues with matching up daughters in the father's family, but three quarters of Catholic French-Canadian girls are named Marie something-or-other, and the names they actual go by are as malleable as their ages, so it can be pretty hard to tell who's who from one census to the next. Plus, they could have gotten married and left home.
  7. Or, this page could belong at the end. It's the only "standalone" page that can't fit between any two of the others.
  8. At least one of the names even matches up to a known St Cyr daughter (Loise = Marie Eloyse). The other one was listed only as Marie in the 1861 census, so she might correspond to just about any of the younger known daughters.

Sunday, July 2, 2017

Taking Up the Canada's 150th Genealogy Challenge

Yesterday was Canada Day, and a very special Canada Day at that – the 150th anniversary of Confederation, the creation of the Dominion of Canada from the former British colonies of Nova Scotia, New Brunswick, and Canada (comprising Quebec and Ontario). Patricia Greber of the My Genealogy Life blog celebrated with a Canada's 150th Genealogy Challenge, asking bloggers to "List all your ancestors that were living in Canada in 1867, the dates they arrived (can be approximate) and where they first settled."

So, albeit a day late, I'm going to answer this challenge for the three-quarters of my maternal lines that are Canadian. (My paternal lines are solidly English, Scots, and German.) My maternal grandmother's father descended from French-Canadian stock who first arrived in Quebec in the late 17th century, while my maternal grandfather immigrated to the U.S. from Prince Edward Island; his PEI forbears came from Scotland (Outer Hebrides) and Ireland.1

All except my 3x great-grandfather Michael Murphy were born in Quebec or Prince Edward Island, so for them there are no "arrival dates" or locations "where first settled". Instead I have listed their probable locations in 1867. (Michael Murphy arrived from Ireland ca 1820 and first settled in Georgetown, PEI.)

AncestorRelationshipProbable residence in 1867
*Louis Rabideau (1850-aft 1913) 2x great-grandfather Saint-Paul-de-Chester, PQ
*Celina Cloutier (ca 1850-1881) 2x great-grandmother unknown
Louis Robidas (1832-1921) 3x great-grandfather Saint-Paul-de-Chester, PQ
Marie Deshaies-St. Cyr (1829-1895) 3x great-grandmother Saint-Paul-de-Chester, PQ
Divine Louise Girardeau( 1790-1875) 4x great-grandmother prob Stoke or St. Camille, PQ
**Jean-Baptiste Deshaies-St. Cyr (1806-unk) 4x great-grandfather possibly Nicolet, PQ
**Victoire Lemire (1811-unk) 4x great-grandmother possibly Nicolet, PQ
Dominic Murphy (1854-1914) great-grandfather Georgetown, PEI
Rose Ann McIntyre (1862-1937) great-grandmother Lot 19, PEI
William Murphy (1830-1909) 2x great-grandfather Georgetown, PEI
Flora Ann McDonald (1832-1911) 2x great-grandmother Georgetown, PEI
Neil McIntyre (1814-aft 1881) 2x great-grandfather Lot 19, PEI
Mary Ann McLellan (1817-1896) 2x great-grandmother Lot 19, PEI
**Michael Murphy (ca 1800-aft 1861) 3x great-grandfather Lot 41, PEI
**Magdelen Morison (ca 1800-aft 1861) 3x great-grandmother Lot 41, PEI

* Louis and Celina married in 1869, probably in New Hampshire, but I believe they were still in Quebec in 1867.
** These ancestors may well have been alive in 1867, but so far I haven't located any death records for them.

  1. PEI didn't actually join the Dominion until 1873, but since it is informally known as the "Cradle of Confederation" for hosting the 1864 Charlottetown Conference where the initial steps toward Confederation were taken, I'm going to stretch the bounds of the challenge a bit to include my PEI lines.

Sunday, April 23, 2017

What Were the Chances? A Tale of DNA

About two and a half years ago, I took an AncestryDNA autosomal DNA test, and linked the results to an Ancestry member tree. I looked over my matches now and then, but practically all of them were in fourth cousin range and beyond, most of them had no tree or only a private one, and few of the ones who did have public trees had any apparent connections to mine. There were a couple of matches around second cousin range and a bare handful at third, but I hadn't gotten around to investigating them further.

Then, a few weeks ago, two new matches, both female, popped up at the very top of my match list – still in 2nd-cousin territory (I have no full first cousins), but the highest yet. The top match (500 cM) was administered by the second match (299 cM), so they looked pretty much like a mother and daughter. And based on one of our shared matches (one of the few which I had been able to identify from their posted tree), the match had to be through my paternal grandfather. Trouble was, the surname was completely unknown to me, and the tree attached to the top match was private.

Even taking into account the possibility of an outlier at the extremes of observed shared DNA ranges, 500 cM is almost certainly in the class that includes first cousins once removed, half-first cousins, great-great-aunts, and half-great-aunts, with a second cousin just barely possible. Anyone at the great-aunt level or older would have been deceased long before atDNA testing began. And, since I have no full first cousins and my grandfather had no siblings who lived to adulthood, I have no (full) first cousins once removed or second cousins on his side.

That leaves half-first cousins. Well, that had potential; before his fifth and final marriage to my grandmother, my grandfather fathered five other children (that I know of) with four different mothers, all before the turn of the 20th century. Thus I had four half-uncles and a half-aunt who could in theory have given me half-first cousins. However, the reality is a bit more restricted in scope:
  • Half-aunt Hazel – four sons with children; one daughter, unmarried and childless
  • Half-uncle Kenneth – married twice, no children
  • Half-uncle Leonard – one adopted son, no biological children
  • Half-uncle Chester – no record after birth registration, assumed died as child
  • Half-uncle Vinal – died as an infant
In other words, my only known female half-first cousin had no children, male or female – and she died in 1995 and was thus certainly not my 500 cM match! Which left, well, no leads at all.


What about half-uncle Chester?

A couple of years ago, I posted some speculations here, entitled Whatever Happened to Chester L. Kirk?, in a 52 Ancestors post. Chester L., the son of my grandfather Chester F. Kirk and Nellie Crosman (who, as it happens, was not one of Chester's five wives), apparently dropped off the face of the earth after his birth was registered. No death record, no census records, no marriage record... he just disappeared. It seemed likely he had died as a child and the record was lost or misfiled.

Research into Nellie's subsequent whereabouts eventually found her in Gloucester, Massachusetts, married to a William Mitchell, with a son Linwood Arthur Mitchell, but no sign of Chester L. However, certain interesting coincidences led me to conjecture that Linwood might actually be Chester, informally adopted and renamed, with a delayed birth certificate naming William Mitchell as his father, and stating a birth date exactly two years (to the day!) after Chester's actual birth.

But, as I said in that post, "Of course, I can't prove (lacking DNA evidence) that any of this is more than just speculation."

What were the chances?

No, there had to be some other explanation. Probably the shared DNA amounts were misleading for some reason (some kind of double-cousin relationship? IBS?) and it really was one of the descendants of my half-aunt Hazel. For that matter, it was hardly out of the question that grandpa could have had additional off-spring that just never came to light before. So I never gave Uncle Chester any serious thought, and with no other apparent possibilities, I started researching Hazel's grandchildren to see if one of them had married into the surname of my new matches.

Two days later, I received an email from a young women who was doing some genealogical research for her grandmother and mother. A couple of months before, she had stumbled across this blog, with its two-year-old post idly speculating about the fate of one Chester L. Kirk. It made for eerie reading, she said, because her family lore held a rumor that her great-grandfather wasn't the biological son of his nominal father.

Her great-grandfather was named Linwood Arthur Mitchell.

Her grandmother – Linwood's daughter – and her mother had just received their DNA test matches.

To a half-first cousin, and a half-first cousin once removed, respectively.

To me.

What were the chances?