Identifying an Unknown Biological Father: A 1920 Case Study

Identifying an unknown biological father from the early twentieth century requires integrating autosomal DNA clustering, relationship probability modeling, and targeted documentary research. By organizing genetic cousins using the Leeds Method and applying What Are the Odds? (WATO) statistical tools, researchers can isolate an ancestor’s likely position on a family tree even when vital records omit a father’s name. Correlating these candidate profiles with local census schedules and more specialized record sets, like military service compensation records, can help confirm (or at least provide strong evidence of) the ancestor’s identity.

Our client, Claire, approached our team to resolve a long-standing family mystery. Her grandmother, Eleanor Ruth Phillips, was born in Pennsylvania on April 26, 1920. Family oral history indicated that Eleanor’s mother, Daisy Turner, was working in Pennsylvania when she became pregnant. Following Eleanor’s birth, Daisy’s sister and brother-in-law, Arthur Phillips, took the infant into their home and raised her under their surname.

No official birth registration recorded Eleanor’s biological father, leaving half of her grandmother’s heritage uncharted. Claire provided our team with access to her genetic data from AncestryDNA, FamilyTreeDNA, and GEDmatch to identify Eleanor’s biological father using genetic genealogy.

Investigating early twentieth-century parentage in Pennsylvania presents distinct historical obstacles. Although Pennsylvania instituted statewide civil birth registration in 1906, compliance varied, and social stigma surrounding unwed mothers frequently caused fathers’ names to be excluded from certificates. Families often arranged informal adoptions within domestic households to maintain privacy. Because these private arrangements avoided legal courts, formal custody paperwork rarely exists in county orphan’s court archives.

Post-World War I industrial mobility created additional complications. Western Pennsylvania, particularly Allegheny County, expanded rapidly as a manufacturing center driven by electrical equipment plants, steel mills, and regional rail lines. Thousands of unmarried laborers migrated from rural midwestern areas into industrial boroughs such as Turtle Creek, residing in boarding houses. These transient lodgers relocated frequently, leaving minimal footprints in decennial censuses. Boarding house landladies often supplied census enumerators with inaccurate birthplaces and ages, compounding the research challenge. Resolving this question required pairing meticulous documentary research with genetic evidence.

Our team began by evaluating Claire’s paternal matches at AncestryDNA. AncestryDNA’s SideView technology segregated her results into maternal and paternal lineages. Paternal ethnicity estimates revealed substantial British Isles heritage and an ancestral journey concentrated in Western and Central Pennsylvania, notably Eastern Allegheny and Westmoreland counties.

To sort hundreds of matches, we applied the Leeds Method, clustering cousins sharing between 30 and 500 centimorgans (cM). Three distinct paternal clusters emerged:

• Cluster B corresponded to maternal relatives of Daisy Turner.
• Cluster C represented descendants of Claire’s documented German grandfather, Paul Bauer (husband of Eleanor Phillips).
• Cluster A formed an unknown genetic group with no connection to either the Turner or Bauer families.

The prominent match in Cluster A was Leonard Morrison, who shared 482 cM with Claire on AncestryDNA and 515 cM on FamilyTreeDNA. According to the Shared cM Project, sharing over 480 cM indicates a close relationship, such as a half-first cousin, first cousin once removed, or half-great-uncle.

Because Leonard shared no DNA with Cluster B or Cluster C, we excluded relationships through Claire’s known grandparent lines. Chronological evidence favored identifying Leonard as a half-great-uncle of Claire, rather than the more distant relationship of great-great-uncle. 

Census and cemetery records established that Leonard and his twin sister, Lois, were born in March 1939 in Turtle Creek, Allegheny County, Pennsylvania. Leonard’s documented father, George K. Morrison, was born in 1910, making him only ten years old when Eleanor was born in 1920. Therefore, George could not be Eleanor’s father. Direct-line Y-DNA testing also indicated that Leonard did not match documented Morrison relatives. These facts revealed that Leonard was very likely not the son of George Morrison, but Eleanor’s paternal half-brother, sharing the same biological father.

Because Leonard was two generations closer to the mystery ancestor, his shared matches provided high analytical value. Cluster A divided into three distinct sub-clusters:

• Sub-cluster 1a descended from James Montgomery Lister (1807–1890) and Jane Sullivan (1807–1894).
• Sub-cluster 1b descended from Wesley Oliver Gresham and Nancy Ann Crawford.
• Sub-cluster 1c descended from James Franklin Stone and Martha Wilcox.

We entered Leonard’s shared centimorgan values into the What Are the Odds? (WATO) tool on DNA Painter to test ten genealogical hypotheses within the Lehman family. Hypotheses outside the lineage of David Stewart Lister (1844–1885), son of James Montgomery Lister and Jane Sullivan, scored 1 or less. The hypothesis placing Leonard’s father as a grandson of David Stewart Lister and Julia Franklin received a score of 15,428.

Documentary research into the descendants of David Stewart Lister and Julia Franklin revealed that while most family members resided in Illinois or moved further west, one grandson relocated to Pennsylvania: Walter Lewis Patterson, born in 1898 to Blanche Lister. We tracked Walter through historical records:

• The 1900 federal census showed two-year-old Walter living with his mother in Bloomington, Illinois.
• A 1918 World War I veteran compensation application documented that Walter enlisted while living in Turtle Creek, Allegheny County, Pennsylvania.
• The 1920 federal census placed Walter as a lodger in Turtle Creek at the time of Eleanor’s conception.
• 1930 & 1940 Records: The 1930 census recorded Walter operating a pharmacy in Turtle Creek. Local records and his wife’s obituary indicate he remained in Allegheny County through the late 1930s, placing him in Turtle Creek when Leonard Morrison and his twin sister Lois were conceived in mid-1938.

Walter Lewis Patterson lived in Turtle Creek in 1919 and likely also in 1938, confirming him as a strong biological father candidate.

Sex chromosome analysis reinforced this conclusion. Leonard shared no X-DNA with Claire, matching biological rules since sons do not inherit X-DNA from their father, and Leonard and Claire’s common ancestor is Leonard’s biological father. Leonard’s Big Y test identified haplogroup R-Y91369, tracing to western Ireland and aligning with Claire’s British Isles genetic admixture.

With Claire and Leonard’s connection to the Lister family confirmed, these findings provide strong evidence to support the identification of Walter Lewis Patterson as the biological father of Eleanor and Leonard. Targeted testing of his descendants could take Walter from a likely candidate to a confirmed ancestor.

DNA clustering and documentary research helped address the question of Eleanor Phillips’s paternity. Walter Lewis Patterson was identified as the likely biological father of both Eleanor Phillips and Leonard Morrison.

Archival records reveal the story of an Illinois native who settled in Western Pennsylvania and remained there following his military discharge in 1919. Establishing himself as a pharmacist in Turtle Creek, Walter lived in the state where Eleanor was conceived in 1919 and the specific community where Leonard was born in 1939.

For Claire, this discovery brought clarity to a family mystery. Her newfound link to the Lister family connected Claire to nineteenth-century Illinois pioneers and deep ancestral roots in the British Isles.