My great-grandfather, George Arthur Walsh (1873-1943), was particularly quiet about his family. When he died, my grandfather was just seven, and it was assumed he had no living relatives. However, when I began my genealogical research, I discovered that he was the youngest of ten children and that two of his siblings survived him.
A few years later, I began using DNA as a research tool and soon discovered a cluster of DNA matches originating in Guysborough, Nova Scotia, where my great-grandfather was both. These DNA matches were closely related to my grandfather and, based on shared matches, appeared to be connected through my second great-granddaughter, James Walsh (1827-1905), but not through my second great-grandmother, Mary Ann Dalton (1830-1921). This discrepancy was peculiar to me, so I began examining this family more closely to understand how they might be related.
Aerial View of Guysborough, NS from Buckley Family Collection at Nova Scotia Archives
One thing I quickly noticed was that based on the family trees of these matches, they all shared descent from a couple: Henry James Gosbee (1875-1926) and Lillian Frances Lucas (1875-1961). Another important detail was that based on the amount of shared DNA between the Gosbee descendants and my grandfather, the only logical conclusion was that either Henry Gosbee or Lillian Lucas was the biological child of James Walsh (1827-1905) —and therefore a half-sibling of George Walsh (1873-1943). Determining which of the two was biologically a Walsh proved tricky, as none of the descendants I reached out to had any information regarding this generation.
The breakthrough came when a Gosbee descendant tested with 23andMe and matched with me. One key difference between 23andMe and AncestryDNA is that 23andMe provides haplogroup assignments, while AncestryDNA does not. A haplogroup is a genetic sequence unique to a specific male or female lineage and can be used to confirm patrilineal or matrilineal descent. This male Gosbee descendant and I did not share matching paternal haplogroups, which meant that Henry James Gosbee (his direct paternal ancestor) could not have biologically been a Walsh. This left only one possibility: Lillian Frances Lucas must have been the biological daughter of James Walsh. Genetically, this conclusion made sense—but when I turned back to the paper-trail, it became confusing.
Lillian Frances Lucas was born on 19 December 1875 in Guysborough, Nova Scotia, the illegitimate daughter of Mary Lucas (1856-1921). On her baptism at St. Ann’s Roman Catholic Church dated 20 June 1876 it is as follows:
“Lilia Francis McDonald, birthdate 19 Dec 1875, father: Ronald McDonald, mother: Mary Lucas, Sponsor: Elibeth Gosbey”
St Ann’s Catholic Church
If Ronald McDonald was recorded as Lillian’s father, why did the DNA suggest James Walsh was instead the father?
One thing to remember when researching is that DNA does not lie, but people often can be misinformed or outright lie when providing information for a number of reasons. In this case, it was likely for two main reasons:
It was also notable that Lillian was not recorded as illegitimate in her baptismal record. This may suggest that Mary Lucas claimed to be married at the time of baptism to avoid her child being labeled illegitimate.
A further search for Mary Lucas in Catholic Church records and Canadian Census records shows that she never married but did end up having two more illegitimate children:
On the baptismal records for both Mary Rose and Mark, no father is named. However, Mark’s baptism, dated 10 September 1887, included an intriguing detail: the father is described as Protestant, which James Walsh was. We do not have DNA from descendants of Mary Rose or Mark, so it cannot be proven or disproven that James Walsh was also their biological father, but the clue on Mark’s baptism offers a little support to the theory that he was.
Whether my great-grandfather was aware of these potential half-siblings remains unknown. All three outlived him: Lillian died in 1961 in Gloucester, Massachusetts; Mary Rose (Gertrude) died in 1949 in Boston, Massachusetts; and Mark died in 1956 in Guysborough, Nova Scotia.
This case illustrates how DNA can meaningfully expand and clarify traditional family history research, particularly when documentary evidence is incomplete or misleading.