I Know This Much (And It's Not a Lot, But I'll Take It)

He stares out at me, his face obscured in a blur of light and shadow, his portrait the victim of an unsteady photographer's hand, grasping for the image of an image behind glass. The glare is a bright warning: you won't get far in getting to know me.  

On the day that I settled in behind the steering wheel and buckled up to cruise the byways of my family tree, the man in the photograph above became my first true roadblock.  

I knew his name:  Thomas J. Barnett.  He was my paternal grandmother's father.  His name was always known, anytime I asked.  Tom Barnett.  That was Grandma's daddy.  He died when she was a little girl.  Grandma was already almost the oldest person I knew.  I guess I did not think to ask her, well, who were Tom Barnett's parents? They were definitely long gone by then.  The truth is, I'm not even sure Grandma would have known the answer.  Now, nearly 30 years after she passed away, the origins of her father and his family seem to be beyond the knowledge of anyone living.

Earlier this year, as the pandemic sidelined us all, I hunkered down and got serious about doing what I could to solve the mystery of my great-grandfather's family using genetic genealogy. 

Using DNA for Clues

The best sources of information on the identity of parents -- birth certificate, death certificate, census records, and obituary -- are either nonexistent or elusive when it comes to Thomas J. Barnett (let's refer to him as TJ, from here on out).  He was born and died before vital records in Arkansas were filed, he doesn't appear in any census records in 1870 or 1880 when he would likely have been with parents, and, if an obituary was even printed at the time of his death, and if said obituary named his parents, it is inaccessible to me, perhaps on inaccessible microfilm in the archives of the Yell County Record or the Dardanelle Post Dispatch.

(For a list of what I do know about TJ Barnett, see the **List of Knowns** at the bottom of this post.)

Because TJ was born either during or not long after the Civil War, I had some thoughts that perhaps he was the product of a soldier's tryst with a local woman, that perhaps she died during or shortly after his birth, and he was raised by her family or was given to neighbors.  This would explain my difficulty in finding him in the 1870 and 1880 federal census records.  (Believe me when I tell you I have scoured all 1870 federal census records in Yell County, Arkansas, for a boy named Thomas or Tom who was living with relatives, had a different surname than the head of household, etc.)

Meanwhile, DNA matches were telling me some truths that were not theory:  Barnett was indeed a biological family surname, but whether it was on TJ's mother's or father's side, I couldn't say. 

I was finding lots of DNA matches with the surname Barnett in their family trees, and many of these matches were descendants of AB and Mary (Brewer) Barnett. To help me sort out this network of matches, I created a research tree, filling the branches with the children of AB and Mary Barnett. I then built the branches downward until I got to the person who was a DNA match with me.

Knowing that I share common ancestors with a group of people isn't enough.  To figure out how TJ Barnett fits into this family, I had to know the meaning of the numbers.  By this, I am talking about the shared centimorgans (cM) between two people with matching segments of DNA.  There are known ranges.  For instance, since the total number of cM on the length of all of your chromosomes is about 7,400, you should expect to share anywhere from 3,600-3,700 cM with each of your parents.  For someone sharing great-great-great-grandparents with you, the numbers are much smaller, but also way more varied.  The average number of shared cM between 4th cousins is 35, but the range could be between 0-135 (see the "Shared cM Project 4.0 tool, v. 4" at https://dnapainter.com/tools/sharedcmv4).  

Thank goodness for the smarties behind the Shared cM Project, as well as others who have carved a path and taken what can be extremely dense and complicated concepts and packaged them into digestible material, some even with how-to steps, like Diahan Southard in her book, "YourDNAGuide: The Book." 

Using Southard's book as a guide, I analyzed the numbers and looked for a generation of connection.  

Before I go on, I have to disclose that I didn't just use my DNA.  I have access to a paternal aunt's DNA.  Being one generation closer to both Grandma and TJ Barnett, her DNA is less-recombined than mine.  That turns out to be a good thing for the Barnett side of my quest.  So, when I mention the generation of connection, it is the generation with my aunt that I set out to identify.

I believe I found the generation of connection in a DNA match sharing 113 cM with my aunt.  This number is really low to be a 2nd cousin relationship, so I moved forward on the premise that this match and my aunt are 3rd cousins.  The match is roughly the same age as my aunt, so I don't think there is a "removed" relationship involved.  Below is how AB and Mary Barnett fit in the match's tree:

 

Who were AB and Mary Barnett, this match's great-great-grandparents?  They were both born in or about 1808.  Some sources put AB and Mary as having been born in North Carolina, while others say AB was born in South Carolina and Mary in Georgia.  They settled in Lauderdale County, Alabama, from at least 1830 through 1840, until a pre-1850 move to Fayette County, Alabama. Their final move was to Franklin County, Arkansas, sometime before 1870.  This is where AB and Mary lived until their passing in 1896 and 1897, respectively. 

AB and Mary had five sons and three daughters born between 1835 and 1856.  So far, I've found DNA matches with descendants of five of the eight children. I wanted to see if any of the children could be excluded or possibly "proven" as a parent of TJ Barnett based on these DNA matches.  

To do this, I utilized a great tool called "What Are the Odds?" (WATO).  Brought to you by the same brilliant folks behind the Shared cM tool, WATO gives you the ability to plug in shared cM numbers with multiple matches to create a probability tree (see link here:  https://dnapainter.com/tools/wato).

Using all of my aunt's matches within this genetic network who shared 40 cM or more (40 is the suggested lowest threshold for WATO - anything lower can skew the results), I created a tree with the descendants of AB and Mary Barnett.  

Let me just say that I love the idea of a probability tree.  It's such a useful tool when all you have to go on are the numbers of shared cM.  However, it's not a magic wand that gives you the name of the correct ancestor in a poof of smoke.  It's still necessary to use available records, common sense, and good old genealogical instincts to draw conclusions.

Back to Records Research:  Yell County

I had researched each of the children of AB and Mary Barnett.  I knew that three sons and two daughters had lived in Yell County, Arkansas.  Oldest son Andrew Franklin had joined the Confederate cause in July 1861, and when he was discharged later that same year due to a medical condition, he listed his residence as Yell County.  Sons Silas Luther and Newton Jasper had both joined the same Union company of the 4th Arkansas Cavalry in Dardanelle in 1864.  Silas would go on to live in Yell County until his death in 1920.  Daughters Sarah Jane and Mary A. both called Yell County home; Sarah was there in the 1870 and 1880 censuses, while Mary A. and her husband-to-be William Flurry were both residents of Yell County according to their marriage record in 1866.

My work was cut out for me.

Andrew Franklin Barnett had no descendants (that I could find).  He had enlisted twice during the Civil War, first with Captain Hallowell's Company of the 15th (Northwest) Arkansas Infantry (McRae's Battalion of Yell County) in 1861, and then again in 1862 with Company B, Hawthorn's Regiment, Arkansas Infantry.  In the 1870 federal census record, he and his younger brother, Silas, resided in the household of William Sears in Dardanelle, AR. The brothers' occupations are noted as workers in a sawmill.  There is no indication that Andrew Franklin (called Franklin in this record) is married.  However, I couldn't discount Andrew Franklin Barnett just yet.  I would come back to him.

Silas Luther likely was married in 1870 when he lived with his brother and worked at the sawmill.  Silas has many descendants.  His first known daughter with his wife, Virginia Polk McGuire, was a daughter named Laura, born about 1867.  Silas in fact played a key role in TJ's life, and because he lived on a farm next to TJ's future wife's family in 1880, as well as having provided the bond for TJ's marriage to my great-grandmother, Sallie Brewer, in 1889, I halfway suspected that Silas was TJ's father.

Another child of AB and Mary Barnett with a connection to Yell County was Sarah Jane, born in 1842, who probably married John Phillips as early as 1860.  Sarah and John Phillips are located in Yell County in the 1870 and 1880 federal census records.  

Then, there is Newton Jasper, born in 1845. He lived in Franklin County, AR, according to the 1870 federal census, but he had been in Yell County with brother Silas when they joined the Union cavalry regiment in 1864.

One more child with a connection to Yell County was daughter Mary A. Barnett, born in or about 1846, who was married there in 1866 to William E. Flurry.

That left three children of AB and Mary Barnett without documented connections to Yell County (that I could find), who nevertheless ended up in Arkansas and who, with the exception of the youngest son, Henry (who was born in 1856 and would have been too young), could potentially be the biological parent of TJ.  One was James M., born in or about 1839, and the other was Rachel Elizabeth, born in 1850.

Drawing Some Conclusions

To recap, I have DNA matches with descendants of five of AB and Mary Barnett's children.  The ones with no proven DNA connection to me (because no known descendants have tested) are Andrew Franklin, Mary A., and Henry.

Henry could be eliminated as a parent because of his age.  I could also potentially eliminate Mary A. as a parent.  Mary A. appears to have died childless sometime between November 1866 and September 1867, when William E. Flurry married Sarah Crabtree in Yell County.  

But wait, you say - couldn't Mary A. have died in childbirth having TJ, who was then given to another Barnett family member to be raised?  

The thing about DNA is that it can put to rest certain theories fairly quickly.  Sarah Crabtree and William Flurry went on to have at least nine children between 1868 and 1884.  My aunt and I share DNA with descendants of this couple, but rather than this pointing to a Barnett connection, it is instead a result of my great-grandmother, Sallie Brewer, being related to Sarah Crabtree. Sallie and Sarah were first cousins (Sallie's father, Joseph, and Sarah's mother, also named Sarah, were siblings). (Note that there is no relation between Joseph Brewer, Sallie's father, and Mary Brewer, AB's wife).  The amount of DNA that my aunt and I share with the test takers who are William and Sarah Flurry's descendants does not support the double cousin sort of relationship that would exist if William Flurry was the father of TJ.  In fact, it's a little on the low side for what it is.

Of course, it's always possible that Mary A. Barnett had TJ out of wedlock, prior to her marriage to William E. Flurry.  We can put a pin in that for later.

Let's get back to the WATO probability tree.  The results of the tree gave me four possible scenarios:

Scenario #1: TJ’s mother was Sarah Jane Barnett.

Scenario #2:  TJ’s mother was Rachel Barnett.

Scenario #3: TJ’s father was Andrew Franklin Barnett.

Scenario #4:  TJ was a son of AB and Mary (Brewer) Barnett.

Based on the hypothetical scores, Scenario #4 is least likely, and the differences between Scenarios #1, #2, and 3 are minimal, although based on the available DNA matches, Scenarios #1 and #2 are scored higher.  

For Scenario #1 or #2 to be true, however, there would need to be a half-relationship because I do not find DNA matches with descendants of either the parents of Sarah Jane’s husband, John Phillips, or those of Rachel Barnett’s husband, Newton Belt.  If true, it would mean that TJ was born out of wedlock or from a previous marriage (in which case, why take the Barnett surname, unless the father was a Barnett relative of Sarah Jane and/or Rachel).

Also, Sarah Jane gave a son born in 1877 the name Thomas, which doesn't make sense if she already had a Thomas in TJ.

My thought is that Scenario #3 would be scored highest if there were actual DNA matches from Andrew Franklin Barnett descendants, but I am of the opinion that TJ was an only child, so there won’t ever be matches outside of those who are TJ’s descendants, which can’t lend any perspective here.

The WATO tree also tells me what is not possible or would be highly unlikely based on the DNA data:

1. James M. Barnett could not be TJ's grandfather.

2. James M. Barnett (who had a son named Thomas in 1886, so why would he have two sons with that name) could not be TJ's father.

3. Silas Barnett could not be TJ’s father.

4. Newton Barnett (who also named a son Thomas in 1879, so again, why name two sons the same name) could not be TJ’s father.

5. It’s highly unlikely TJ was a son of AB and Mary (Brewer) Barnett).

Here's my takeaway from all of this. There are a lot of assumptions/leaps that have to be made in order to make it work that a daughter of AB and Mary Barnett was the parent of TJ, not the least of which is the idea that he was born out of wedlock to one of three women who ended up marrying a man other than TJ's father. I think the simplest and most likely scenario is the one in which Andrew Franklin Barnett was the father of TJ.

Other than what I was able to find about the military career of Andrew Franklin Barnett (let's call him AF), I knew that he was married briefly to Rachel Rose before his death in 1886. Rachel was at the time of her marriage to AF twice widowed, and she would go on to marry my great-great-grandfather, Joseph Brewer, after AF's death.

Joseph Brewer was a Union veteran, and when I obtained his Civil War pension application file a few years ago, I gained more documentation involving TJ. Contained in Rachel (Rose) Brewer's application for a widow's pension when Joseph died in 1898, there is an affidavit signed by TJ stating that he had known Rachel fifteen years. Rachel's second husband, Thomas Wadsworth, had apparently died in 1881. Rachel and AF were married for about five years, according to Rachel.

Excerpt of Rachel (Rose) Brewer's affidavit filed with her application for widow's pension, 11 Jul 1899, upon the death of fourth husband, Joseph Brewer

I contacted the Arkansas State Archives to see if there was any record of AF filing for a pension during his lifetime, or Rachel, or even AF’s mother, upon AF's death in 1886 or in 1891 when the law governing pensions was expanded, but there is no record appearing in the official rolls for an application by or for Andrew Franklin Barnett or AF Barnett.

Like many ancestors, AF is proving to be somewhat elusive. I feel lucky to have as much information as I do. What would be most helpful is to find a marriage record or other source linking him to TJ's mother. Of course, it would help if I knew her identity! Should be simple, right? Just use genetic genealogy!

I am working on it! Yet, even with all of the nifty tools and books available, this is harder than it looks. (I didn’t get into all the details of the work I put in researching TJ's family for brevity's sake.) Figuring out TJ's mother's identity is proving to be a challenge and a half! All I can say so far, based on deductions made using the same methodology above, is that I believe TJ's mother was a Smith.

Gee. That's promising, isn't it?

The main issue seems to be that my aunt just did not get as big a dose of the Smith DNA as she did Barnett. In some instances, I share greater amounts of cM with our Smith matches than she does.

I plan to return to this at a later time, perhaps when there are better DNA matches (positive thinking!), or after I've taken a research trip to parts of Mississippi and Alabama to more fully explore records. For now, I'm happy to claim AB and Mary (Brewer) Barnett as my 3rd-great-grandparents. It reduces the glare surrounding TJ a bit ... or is that just my imagination?

List of Knowns - Thomas J. Barnett

• He was born in Arkansas sometime between 1864 and 1868, according to the only two census records in which I have located him (1900 and 1910 federal censuses).
• He married my great-grandmother, Sallie Brewer, in 1889 in Yell County, Arkansas, and his marriage bondsmen were S.L. Barnett and H.M. Christy.
• He lived in Prairie, Yell County, Arkansas, in 1900, occupation farmer, with his wife, Sarah A. (Sallie), and sons Ernest and Silas (my grandmother's brother who went by Bill).
• He lived in Yell County, Arkansas, in 1910, still a farmer, with his wife, Sallie, and children: Ernest, Silas (Bill), Ollie, Esther, Etta (listed as Etter).
• He likely died between 1911, when another daughter was born, and 1914, when son Ernest's WWI draft document states that he (Ernest) is supporting his mother and three of his sisters.

Besides the records mentioned above, TJ appears in the following items I've collected:  

• Signed attestation to his wife's former stepmother's 1899 Declaration for Widow's Pension, stating that he had known the applicant for 15 years;
• His widow Sallie's death certificate;
• The delayed birth certificates of sons Ernest and Bill and daughter Ollie;
• His daughter Esther's Social Security application; and
• His son Ernest's obituary.

Source Citations

Select Records - TJ Barnett

1900; Census Place: Prairie, Yell, Arkansas; Page: 4; Enumeration District: 0158; FHL microfilm: 1240080

1910; Census Place: Prairie, Yell, Arkansas; Roll: T624_68; Page: 2A; Enumeration District: 0172; FHL microfilm: 1374081

Arkansas Marriages, 1837-1944, FHL: Film #007579405

U.S. WWI Draft Registration, Registration State: Arkansas; Registration County: Yell; Roll: 1530658 

Select Records - AB and Mary (Brewer) Barnett

1830; Census Place: Lauderdale, Alabama; Series: M19; Roll: 1; Page: 192; Family History Library Film: 0002328

1840; Census Place: Lauderdale, Alabama; Roll: 7; Page: 79; Family History Library Film: 0002333

1850; Census Place: Beat 5 Division 16, Fayette, Alabama; Roll: 5; Page: 22A

1870; Census Place: White Oak, Franklin, Arkansas; Roll: M593_53; Page: 109B; Family History Library Film: 545552

1880; Census Place: White Oak, Franklin, Arkansas; Roll: 44; Page: 735A; Enumeration District: 059

Original data: Find A Grave. Find A Grave. http://www.findagrave.com/cgi-bin/fg.cgi.

Select Records - Andrew Franklin Barnett

Compiled Service Records of Confederate Soldiers Who Served in Organizations from the State of Arkansas published by NARA

1870; Census Place: Dardanelle, Yell, Arkansas; Roll: M593_67; Page: 538B; Family History Library Film: 545566

The National Archives at Washington, D.C.; Washington, D.C.; Record Group Title: Records of the Department of Veterans Affairs, 1773 - 2007; Record Group Number: 15; Series Title: U.S., Civil War Pension Index: General Index to Pension Files, 1861-1934; Series Number: T288

Select Records - Silas Luther Barnett

1870; Census Place: Dardanelle, Yell, Arkansas; Roll: M593_67; Page: 538B; Family History Library Film: 545566

1880; Census Place: Magazine, Yell, Arkansas; Roll: 60; Page: 348C; Enumeration District: 219

Index to Compiled Military Service Records; 4th Arkansas Union Cavalry - Arkansas Research, Historical Data Systems, Inc.; Duxbury, MA 02331; American Civil War Research Database

Select Records - Sarah Jane Barnett

1870; Census Place: Delaware, Yell, Arkansas; Roll: M593_67; Page: 547B; Family History Library Film: 545566

1880; Census Place: Magazine, Yell, Arkansas; Roll: 60; Page: 348C; Enumeration District: 219

Select Records - Newton Jasper Barnett

Index to Compiled Military Service Records; 4th Arkansas Union Cavalry - Arkansas Research, Historical Data Systems, Inc.; Duxbury, MA 02331; American Civil War Research Database

1870; Census Place: White Oak, Franklin, Arkansas; Roll: M593_53; Page: 109B; Family History Library Film: 545552

Select Records - Mary A. Barnett

"Arkansas County Marriages, 1838–1957." Index. FamilySearch, Salt Lake City, Utah, 2009, 2011.