The Strange Case of the Traveling Barber

In previous episodes of Genetic Mysteries, I gently pushed over some weedy brick walls in my genealogical garden, when I worked out the ancestry of my great great grandparents, Elizabeth Lovell and her husband Hugh Humphrey (fine pieces of sleuthing I might add). See the articles, The Ancestry of Elizabeth Lovell and Reconstructing Hugh Humphrey for the finer details. A dozen years on, however, we now have readily available at-home DNA testing, which allows us to shed new light on a variety of what were once genealogical mysteries.

I spit into my first DNA bottle-cap just about 10 years ago, and sent it off to Ancestry.com with anticipation of what I might discover. Turns out my parents were faithful observers of their marriage vows, by which I mean that no half-siblings have come to light (yet). Over the last decade, I've been diligently tracking down DNA matches, and plugging them into my family tree in their rightful places. All the pieces seem to fit nicely, but taking the broad view, as you can see in the pedigree chart below, there is an obvious and glaring hole. None of David Henry Humphrey's ancestors show up in my DNA matches.

Let me start with a little background on what this image represents and some useful terminology starting with this simple match (Ancestry ThruLine™) to David Humphrey's wife.

When discovering a new DNA match that shares a significant amount of DNA with myself, I add them to my database along with any of their ancestors that descend from our shared most recent common ancestor (MRCA). In the example shown here, Mr. H. and I both descend most recently from Phoebe Jane Compton. If we mentally complete the lines shown in the diagram (our independent lines of ascent), you can see that those lines must pass through each ancestor, merging into a single point at the MRCA. If I were to find an additional match to his sibling later (see the image below), I could use the same process to traverse each of those ancestors a total of two times each, that is, except for Mr. H. and his sibling, which were only traversed one time each. This is the DNA traversal count. This count represents the number of DNA matches that support an ancestor's biological relationship. In the pedigree chart shown above, the green boxes indicate that the ancestor was traversed, and the number in the box indicates how many times. But, you might ask, what about Phoebe Compton's husband, should he not also be a match to Mr. H. and myself? The answer is most assuredly, no! The purpose of using DNA in the pedigree charts is to show proven biological lines of ascent. We cannot be certain that both of her children had the same father; they may be half-siblings.

So, as mentioned above, I have not located a single DNA match that descends from the parents of my great grandfather, David Henry Humphrey. I'm assuming here that he is the biological father of Minnie Elizabeth Humphrey, but without matches to his ancestors, this cannot be shown conclusively. I've found matches at all the major websites: Ancestry.com, MyHeritage, 23andMe, FTDNA, and GEDMatch, and although some of the listed websites import Ancestry.com's raw data, MyHeritage and 23andMe use completely independent DNA samples. I have found matches for my father's other ancestors for these same generations and beyond, with the exception of the unverified males shown, Alpha Forsyth, Edmund Scully and William Godsil. These, and all my mother's ancestors, will be shown presently, to be unrelated to this case.

So now that we know a mystery is afoot, there are several possibilities that we might consider. The first of these, and the most obvious, is the question as to whether David Henry Humphrey might have been adopted. Let's take a look at what we know about his birth. His tombstone [1] indicates that he was born in 1856, but tombstone order forms are seldom filled in by the deceased. So turning to the 1900 U.S. Federal Census [2], the only one which lists partial birth dates, he, or more likely his wife, listed his birth as March of 1856. His daughter, in her, A Book of Memories: Vol. 1 [3], wrote the following.

David Henry Humphrey, son of Hugh Humphrey and Elizabeth Lovell, was born in Vermont, McDonough County, Illinois, March 16, 1856. His father was born in Great Barrington, Mass., in 1819. There is no record of his mother's birthplace, but she was married to his father in Cincinnati, Ohio, in 1844. David had one sister, Susan Alwilda, and a brother Francis Marion, His brother died at about eighteen years of age.

David Humphrey was a barber by trade, and opened his first of many barber shops in Avon, Illinois, where he met his wife and started a family. From there he moved to Maquon, Illinois and then to Galesburg. When he arrived in Galesburg, he did so while walking down Main Street with his barber chair strapped to his back, and I like to think, whistling, Yankee Doodle. He operated barber shops in both Brown's and the Union Hotel. It was there that Carl Sandburg, poet extraordinaire, as a young boy worked for him sweeping floors. Carl remembered him fondly in his autobiographies, Always the Young Strangers and Prairie Town Boy. This was my great grandfather.

David's parents, Hugh Humphrey and Elizabeth Lovell are verified by Hugh Humphrey's obituary [4] where it states

Hugh Humphrey was born in Great Barrington, Mass., in the year 1819, where he resided until at the age of 21, when he went to Ohio and settled in Cincinnati, Hamilton, County, and in 1844, he was married to Miss Elizabeth Lovell. In 1856 he came to the state of Illinois with his family, and settled in the south part of McDonough County where he resided until 1868, when he moved to Cass County, Mo., and in 1874 he returned to Prairie City, Illinois where he has lived until his death which occurred Feb. 2, 1890. ... He leaves a wife and two children to morn his loss, Mrs. A. P. Estus, of this place and D. H. Humphrey of Maquon, Ill.

David (D.H.) was born in the same year that, according to his daughter, the family moved [5] from Anderson Township, Hamilton County, Ohio, located just East and outside of Cincinnati, to McDonough County, Illinois. His parents were married in Cincinnati in May of 1843 [6], and they had their first child, Levi, around 1845. There is no record of his birth, but he appears in the 1850 U.S. Federal Census [7] as 5 years old. He must have died young, because he does not appear in the 1860 U.S. Federal Census [8]. Their second child, Susan Alwilda Humphrey was born on May 17, 1846, and grew up to marry [9] Andrew Perry Estus. She is the Mrs. A. P. Estus mentioned in Hugh's obituary. My grandmother stated that they had another son, Francis Marion, that died at about 18 years of age. If this were true, he would have appeared in at least one of the census records, which occur every 10 years, but he is not found. One of Elizabeth Lovell's step-brothers appears in the 1850 U.S. Federal Census for her father, David Lovell, and his second wife as, Francis M. Lovell [10], age 10. It is not known when he died. One of Liz's other step-brothers, Levi, is found in 1880 working in the livery of the household of a 27 year old Francis, last name Marion [ 11 ]. These may be mere coincidences, or perhaps one of these young men lived with the family for a short while, and my grandmother confused this person as a child for Hugh and Elizabeth. My grandmother did not mentioned the young Levi. In her defense, she was born after both Hugh and Elizabeth had died, and did not have the advantage of the Internet.

Getting back to the case at hand, we must ask ourselves why they might have waited 10 years after Susan Alwilda was born before having David. There are plenty of normal reasons for this to occur: an unplanned pregnancy, baby envy, deceased children missing from the records, or perhaps an unspoken adoption.

Although I have been able to correctly place a large number of DNA matches into my tree, there are as many, or more, that remain unplaced. Part of the reason for this is that I usually discard matches with shared DNA segment lengths below 20 centimorgans (cM). I only use smaller lengths for more distant relatives for which longer lengths are not possible or are infrequent, or when that is all that is available. Also, many of my DNA matches have not provided ancestry information on the websites where their DNA data resides. This makes it difficult to find out who they are, and where to place them. That being said, I have had a good deal of success locating some of their names on public-facing people-search websites and as children and grandchildren listed in online and newspaper obituaries. Once found, their deceased ancestors can usually be located using the search functions at both Ancestry.com and FamilySearch.org. From there it is a matter of working through their trees until I've found our common ancestor, always being aware that everyone has multiple lines of ascent that sometimes involve endogamy (the intermarrying of closely related relatives), and that online trees are often invalid in one or more generations. It sometimes takes days to fix mistakes and then to pin down a unique common ancestor. Just as often, however, when their ancestors appear to be valid on paper, they still do not overlap with mine. In those cases, I assume there was some illicit canoodling going on, or perhaps an adoption, but where and with whom is usually outside my purview.

But every so often, I catch a lucky break, as it was for the case in question when I discovered that several of my unplaced DNA connections all descended from the same couple. This led to additional searching and tracing until several months later, I had found well over 80 DNA matches who positively descend from this couple with an additional 20 matches that descend from their direct ancestors. I've also found an equal number who share DNA with those same matches, but who still remain unplaced. I've located these matches on all 5 of the testing platforms mentioned above, and descending from 9 of their 14 children, making this a very significant discovery that cannot be ignored or explained away. The closest of these DNA matches shares 1.44% (108 cM) of their DNA with me, placing them firmly between full (or half) 2nd and 3rd cousins, and either directly, or once removed. These predictions are based on statistical calculations derived from large numbers of known relationships, and can be easily determined using standard DNA relationship prediction calculators, Orogen [ ? ] and The Shared cM Project 4.0 tool v4 [ ? ] being but two. The match in question is the same generation as I, so we are almost certainly half 3rd cousins, indicating that this match and I probably share one great-great grandparent, they being the mother or father of David Henry Humphrey.

These new ancestors are Russell Loomis [LH1X-LHT] [ ? ], who was born on August 5, 1769 [12] in East Windsor, Hartford, Connecticut and Corinna Gillett [KG33-V9R] who, according to her tombstone [ 13 ], was born about 1772. They married in East Windsor on October 20, 1791 [14] and had 14 children. Their first child was born in 1792 in Connecticut, and their youngest in Ohio, in 1814. The family was still in East Windsor in the summer of 1810 [15], but by June, 1812, they were living in Windsor, Ashtabula, Ohio when and where one of their sons was born. Their oldest grandchild was also born in 1814 and their youngest, long after David Humphrey was born. If we consider that most children are given up for adoption when their mothers are young, prior to their first marriage, then we can expect David's mother to have been born between 1838 and 1840, the father could have been a good deal older. Reviewing the closest match mentioned previously, they were born in Kingman, Maricopa County, Arizona, and only one of the Loomis’s descendants was living there at that time of their birth. Based on that highly convenient fact (which took many days to track down and isolate), I am able to now positively identify our closest shared ancestor as Sircellaeus Julius Webster [L6R7-ZDF], the son of Mary Ann Loomis [9F3S-KPS] and grandson of Russell Loomis and Corinna Gillett, making us half 3rd cousins, exactly as predicted. According to the 1900 U.S. Federal Census [ 16 ], Sircellaeus was born in Ohio in November of 1838. His father is found in the Ohio tax records [ 17 ] in Kingsville, Ashtabula, Ohio, the seat of the Loomis family, in 1837, so presumably Sircellaeus was born there. By 1850, however they were living in Bradford, Rock County, Wisconsin [ 18 ], and on September 20, 1853, M. W. Webster, acting as minister, is found officiating the wedding between his daughter Cynthia J. Webster and Henry Smith, in Boone County, Illinois, which is just across the state line from Rock County, Wisconsin [ 19 ]. In 1859, Sircellaeus married Harriet Helm in nearby Wiota, Lafayette County, Wisconsin [ 20 ]. So in June of 1855, when David Humphrey was conceived, Sircellaeus was a mere 16 (and a half) years old and probably still living in Illinois where David was born in 1856. It should be noted here that both the Webster and Humphrey families were on the move in the mid 1850s, both travelling from Ohio to Illinois. Granted they lived in different parts of those states, but it does not seem beyond reason that their paths may have crossed at some point.

The following image shows this close relationship.

This next image shows some of the additional DNA matches, descending from Russell Loomis and Corinna Gillett. Less significant matches have been omitted as have those descending from their daughter Corrina Lucinda Loomis. Corrina and her sister Mary Ann Loomis married the brothers Webster, so including her in the database leads to a pedigree collapse in their parent’s generation, and therefore cannot be used as proof for determining their descendants’ biological relationship, as there is no way to determine which line of ancestors provided the DNA.

Returning to the issue of my other unverified male ancestors mentioned previously, we can rule out the Irish emigrants, Edmund Scully and William Godsil, as possible placement for the Loomis family, because neither of them had yet immigrated to the U.S. prior to when David was born. As far as Alpha Forsyth is concerned, he was born on June 5, 1819. According to Alpha's granddaughter's letters [21] to my grandmother, he was a twin, and that date is listed on the tombstone of his sister, Eliza (Forsyth) Mustain, and roughly agrees with his various census records [22], which also indicate that he was born in Kentucky. His family moved to Hancock County, Illinois in 1832 [23] when he was 13 years old. It seems very unlikely that in early 1838, when he was still just 18 years old and working his mother’s farm in Illinois that he could have fathered Sircellaeus Webster who was born in Ohio. I was able to verify using the tools at 23andMe that several of the descendants of Minnie Humphrey’s husband’s sister, Edna Margaret Forsyth, what would be 4th cousins to the aforementioned match, do not share any DNA, but this being a negative result, does not in itself prove anything. Due to DNA recombination, shared segments of DNA in significant amounts do not always remain intact and detectable in distant relations. That being said, it would be surprising if multiple 4th cousins do not match. I am still trying to determine if either of my DNA matches that are isolated to the Humphrey line, also have overlapping matches with the Loomis line, but since they are not on 23andMe or MyHeritage, and since Ancestry.com does not provide the same capabilities for analyzing matches, I will have to wait for further info. When I do, and if so, then Alpha Forsyth can be completely ruled out. Onto my mother's ancestors, I have been able to verify that several other relatives on my father’s side do positively match other Loomis descendants through several of their children, so we can completely eliminate my mother's ancestors for a Loomis placement.

So with that, an updated pedigree for this proposed relationship, sparse as it is, can be presented. In the below image you can tap the image, or drag the slider, to view the before and after effect (if an image does not appear below, please refresh the page).

Now that we've established that David Humphrey was very likely adopted, let's look at scenarios where he wasn't. David's documented mother, Elizabeth Lovell, according to the data taken from her tombstone by an unknown researcher, died on June 26, 1890, at the age of 63 years, 3 months and 22 days. This would make her birthdate March 4, 1827. She is supposed to have been buried in Prairie City Cemetery, in Prairie City, McDonough County, Illinois [ 24 ], but I have not been able to verify this, as her tombstone has not been located by those who have searched the cemetery for it (although another unrelated Elizabeth Lovell has been found there). Going on the assumption that the information is valid, and setting aside that she was born only 2 months after her parents marriage [25], which would be, to say the least, an unusual situation, it seems very unlikely that the 28 year old Elizabeth Lovell who was married and had a nine year old daughter at home, living on a farm in rural central Ohio, could have had an affair with the 16 (and a half) year old Sircellaeus Webster who was living in Wisconsin at the time.

The next scenario to consider is that David Henry Humphrey is not the biological father of his daughter, Minnie Humphrey. As we demonstrated in the first DNA chart shown above, David’s wife, Phoebe Jane Compton, is the mother of Minnie Humphrey, and her brother, Herschel. We also know this through DNA matches to well over 50 DNA connections to Phoebe’s mother, Julia Ann Taylor. However, as I mentioned earlier, without any matches to David Humphrey’s ancestors, we cannot be certain he is her father. Let’s look at whether or not one of the Loomis descendants, namely Sircellaeus’s son, Leo Arlo Webster [MBMX-5MB], the most direct line, could be Minnie’s father. If so, it would make the closest match and myself 2nd half cousins, which actually has a higher probability due to the large amount of DNA that we share and that I share with their child. In the 1900 U.S. Federal Census [26], for Galesburg, Knox County, Illinois, her birth is listed as December of 1894, Illinois. In her A Book of Memories: Vol. 1 [27] she listed her birth as happening in Galesburg on December 26, 1894. According to both the 1880 U.S. Federal Census [ 28 ] and the 1900 U.S. Federal Census [ 29 ], the 15 year old Leo Webster was living in Fayette County, Wisconsin at the time that Minnie Humphrey was conceived, so once again we are looking at a near impossible situation. Leo Arlo Webster had a twin brother and two older brothers who were 12 and 19 years older than he. Both of the older boys were married, had children, and like their younger brother were living in Wisconsin. The twin brother was as well. It would appear that there are no other reasonable scenarios other than David Humphrey being a child of Sircellaeus Webster by an unknown non-marital union.

There is another part to this case that I have yet to mention. David Humphrey's sister, Susan (Humphrey) Estus, had 5 children, one of which died young, and one who never married. Of the three remaining daughters, all married and at least two of them had children. So where are they now? I do not have DNA matches to any descendants of Susan (Humphrey) Estus, which if any of them had done a DNA test on any of the websites mentioned, and if she and David were biologically related, I would expect to find a match. Since I have not, this is further support for the argument that they may not be biologically related. That being said, it is unclear how many living descendants we can expect Susan (Humphrey) Estus to have had, and they sure do look a lot like.

There is one last point worth mentioning that supports my claim. I am in contact with a Lovell descendant, D. Lovell, who has done their DNA testing with Ancestry.com. Their results include over 20 DNA matches to other members of the Lovell family, including to their 5th cousin, J. Terry, with 23 cM shared between them in two segments. D. Lovell should be my 5th cousin, and J. Terry my 4th, with Elizabeth Lovell, my great great grandmother, being a full sister of the Henry Lovell (1836-1865) shown in the below image, and since I should be related to J. Terry one generation closer than D. Lovell is, I would expect to share more DNA with them. However, neither D. Lovell or J. Terry appear in my Ancestry.com match list, indicating that we do not share DNA in a significantly measureable amount. I also do not match any of the other 20 individuals who match D. Lovell, each sharing 20cM or greater with them. This is probably because I am indeed, not biologically related to the Lovell family. It could also be due to DNA recombination occuring in my most recent ancestors, but that is probably a stretch to consider, though not impossible.

Until further information comes to light, it would seem to be the case of highest likelihood that David Henry Humphrey was born in Illinois and given up for adoption soon after. It is not known why Hugh Humphrey and Elizabeth Lovell left Ohio for Illinois, nor where or how they came to adopt the infant David Humphrey. And since David's daughter seemed blissfully unaware of this adoption, I assume that he was also. All that is left to prove beyond any reasonable doubt that this Loomis-Webster connection is correct, is to find among my still unplaced DNA matches, a female ancestor, shared among many, who was born about the same time as Sircellaeus, and who lived in Northern Illinois or Southern Wisconsin in 1855. This will only be possible if she is in turn not related to Sircellaeus. This elusive young lady might well be David Humphrey's biological mother.

Setting aside these minor complaints, it would appear that David Henry Humphrey’s relationship to the Loomis family in The Strange Case of the Traveling Barber, has finally been ...

---