A comparison of ethnicity results and why you should build a family tree.

This is my most recent admixture breakdown from Ancestry.com. On the right hand side you can see the areas that are “no longer in estimate.”  It’s worth noting that DNA does not recognize geographic boundaries, only DNA testing companies try to do that – an issue which often leaves people puzzling over their results. I’m not at all surprised to see that I no longer have 5% of my ethnicity under the Finland/Northwest Russia region. It was even less likely that I would connect in any way to the Iberian Peninsula so that is not a loss either. I would suspect that my “Ireland/Scotland/Wales” category is mainly Scottish DNA and I would be surprised if Ancestry ever broke this category out for me and started suggesting specific places in Ireland where my ancestors came from.
Living DNA analysis.

In 2016 I tested with the (then) very new LivingDNA company (graphic above). Their specialty is to isolate areas in the UK where your ancestors may have come from. It’s interesting to look at their map of suggested regions and see how my Ancestry DNA results appear to be converging around the same areas. From my paper trail I do know that the “top” areas in my Living DNA breakdown – Northumbria (27.5%), Central England (25.7%) and Aberdeenshire (11.9%) are representative of many of my known ancestors. In a breakdown of my deeper ancestry through Living DNA (graphic below), their chart encompasses most of the continental European ancestry that is suggested by Ancestry’s graphic.

Living DNA's analysis of where my ancestors may have lived 500 years ago.

So for me at least, the ethnicity estimates may be aligning. In this round of Ancestry updates I have heard various anecdotal complaints about the entire percentage of southern European, French or Italian ancestry “disappearing” from an estimate. It’s important to read all of the information that Ancestry provides when they make an update, which you will usually find on your DNA story landing page. This time around they have managed to address questions that may be specific to your update in a box beneath the new estimate percentages. There is even a thumbs up, thumbs down instant feedback capability, and they have developed a sense of humor in addressing some of the disappointment testers may face on receiving the new update. Noting that I had lost a small percentage of Iberian peninsula, they suggest that a common question might be “With my Iberian Peninsula region gone, should I quit matador school?” (Although the more serious question is answered in “What happened to my other regions?”).

As always with ethnicity estimates, updates and discrepancies from company to company, the important things to remember are:

  • This is an estimate of where your ancestors may have come from. The science is not currently advanced enough to “pinpoint” a location reliably beyond the continental level
  • Read the information that each testing company provides about how they test. Ancestry, for example, runs your sample for each region 40 times and gives an average percentage for each one. Each company has its own algorithms and reference panels. Your results will differ from company to company.
  • The relative matching features of testing companies are very accurate, although fewer people are interested in this as an outcome. If you are on Ancestry, consider adding a basic family tree to your profile and be open to people contacting you asking how you may be related. It may not matter to you, but you could be helping someone in the search for their biological family.

Philip Holmes – why every little match matters.

The last post on the search for Philip Holmes was, well, on the technical side. But, it served to illustrate a point. Finding family through DNA matching is not effortless fun. There is a huge learning curve at the start to figure out new language and resources to help understand what you are seeing when you open up your DNA results. What are centimorgans and why is the shared DNA at companies sometimes shown in percentages? Are segments important? Why does it say that a person is your second cousin when I know they are not?

It’s usually necessary to drill down through your match list to make something meaningful out of it. Smaller details such as which of your matches also match each other can be incredibly useful information. New techniques and tools for sorting matches appear with some regularity. There’s constant learning to be done.

How the X chromosome opened up this search.

You can’t see X chromosome matches at every site, but you can on Gedmatch, if you choose to upload there. Usually X matches are small and won’t help establish how you are related. However, when a female matches a male on X, it can feel a little like winning the lottery. This is what happened when Carol’s mom matched with someone on Gedmatch at around 40cMs, including a decent chunk of X (in this case, only 17cMs which some people would barely consider of the threshold of a decent match.) It meant she could only be related through that male’s mother’s side of the family. If we could get information from the match about their family tree, this could be an interesting development.

Luckily, Carol has a high success rate in contacting matches and asking them for information. It so happened that the match’s mother was the family genealogist. She was intrigued with the details of Carol’s search for Philip Holmes. Although she was pretty sure that the name Holmes was not represented in her direct ancestors, she gave us enough information for us to help build an accurate tree back to her 3rd great grandparents. From the size of the match, this is where we thought the connection should be.

At this point, the beauty of the X match kicked in. Based on its unique inheritance pattern, we knew that there were a number of her ancestors that we could discount. (Well, technically her son’s ancestors as he was the match). At great-great grandparent level, it meant there were only 5 surnames to consider. Even at great-great-great grandparent level, there were only 8 surnames that had to be consider. Given that we each have 32 great-great-great grandparents, that is a significant paring down of the possibilities.

To try and keep things straight, I downloaded a ridiculous-looking graphic from the internet so that I could mark the ancestors whose names we had to consider. I’ve edited the tree to protect the identity of the match, but this is what it looked like at the 3rd and 4th great-grandparent levels when I had completed that.

Boxall tree with x symbols for blog

I ran all of these surnames through the surname search on Ancestry but there really wasn’t any consistency to point me in a particular direction. This was hindered by the fact that many of Carol’s mum’s matches were distant and/or had no trees. It was frustrating, but I felt determined that somehow this match was going to give us a lead.

I had already made a few trees for a few Philip Holmes “potentials” – men who were around the right age to have been Carol’s grandfather and who were based in London. I poured over those too, looking for any of the surnames from the match tree. Again, there were none. So what now?

The answer proved to be this: I had to go back to basics. 

DNA evidence on its own does not constitute genealogical proof. Sometimes it may be all that is available to show a connection between two people. But that conclusion should only be made after a thorough search of related records.

I started reading through the records attached to one of the Philip Holmes possibilities and soon made two discoveries.

  • The ancestry for his maternal line contained a possible error that we had been unable to solve. His mother, Kate Phillips, was listed as the grandchild of a Mr and Mrs William Dobney  in the 1871 census. In the tree we had added her parents as Henry Phillips and Sarah Dobney, based on that census record. There were no other records linking her to the Dobney name, so the ancestral line was stuck there.
  • A close reading of the 1911 census record for Philip Holmes, (age 7), living with his parents in Hackney, showed that under the same roof there was an interesting occupant.  Emily Tilbury, aunt (relationship to head of household), aged 74.

And where had we seen that surname before? Right there in the tree with the daft X graphic.

Tilbury only tree

So, if we could connect that Ms Tilbury to the Tilburys in the match’s tree then surely, surely we had found our man.


The point at which I became a DNA geek.

The next part of the search for Philip Holmes involves a little bit of geekiness, and a slight detour.

At some point in the learning curve of using DNA matching in genealogical searches, I believe some of us cross the line from “enthusiast” to “geek”. For me I think it came when I discovered what is referred to as “the unique inheritance pattern” of the X chromosome.

The what?

Exactly. Not the kind of sentence I thought I would be writing when I first started delving into genealogy. Now it flows from the keyboard so naturally that I know I have moved into a different realm of understanding. I might even admit that I scare myself, ever so slightly, when I use phrases like that. I am not a scientist, a biologist or a geneticist, nor do I claim to be. In college I “majored” in English literature and language, which is probably why I can explain literary terms such as “trochee”, “synecdoche” or “enjambment” – (and which also likely confirms my place in geekdom).

In a nutshell – we have 23 pairs of chromosomes (and that is one reason why there is a testing company called “23andMe”). The first 22 pairs are referred to as “autosomes”. The 23rd pair also have a scientific name but are more easily explained by calling them “sex” chromosomes. The chromosomes come in pairs because we inherit one from each parent. The 23rd pair will determine if we are male or female.

X plus X = female.

X plus Y = male.

For various reasons, genealogists researching DNA matches don’t often use matches on the X. But in certain cases it can be significant and exciting because of the aforementioned “unique inheritance pattern.”

The what (again)?

In an even smaller nutshell, (within the first nutshell?)…this:

X DNA inheritance

Females get an X chromosome from each parent, but males only get X from their mothers. This means, if you have a large enough DNA match with a male on the X chromosome, and you also share some autosomal DNA with the match, you have to be connected through his maternal side. Because X can’t be passed down from one male to another, that automatically excludes certain lines of his ancestors from your search.  You can only be related to a male match on the X chromosome through one of his ancestors marked in green in this chart.

X DNA male inheritance.

(This chart and the one before it were taken from the genie1.co.au website written by Louise Coakley. )

We can sit and wait for days or months on end, fiddling with the trees of 4th cousin matches, working down from their great, great, great grandparents through the generations, with collateral lines increasing exponentially as we go. Then one day, we wake up, check our DNA matches and discover that the one match that we have long sought to break the mystery has finally appeared. In this case it was a fairly measly-looking potential 4th cousin match that turned up on Gedmatch. He shared both autosomal and X DNA with Carol’s mum. Importantly, he also did not seem to be a matching kit for what we called “the usual suspects” in this search – a large group of people we knew to be connected to Carol’s mum’s side of the family. And because this match was male, it meant that potentially we would only have to consider that small selection of ancestors in green in the chart. Of course we would still have to work out if he connected to any ancestor by the name of Holmes in London at the beginning of the 20th century. How hard could it be?!

Philip Holmes – one man and his mule.

To go to the beginning of this story, click here. For the preceding installment about the information on the back of the picture, click here.

Photographic “evidence” part 2.

Sometimes when you are on the trail of a missing ancestor, you spend a lot of time messing about with trivia and tiny details, theorizing about a person’s life based on anything tangible that remains in your possession. Often this boils down to photographs without names, dates or locations marked on them.

The photograph of Philip Holmes was systematically dissected, element by element, in the hunt for any clue about his life. When all you have is a photograph of someone and no other information about what sort of life they lived, it is truly amazing how many lines of inquiry you can follow up – even when many of them may be dead ends.

For example: here is part of the photograph of Philip with red circles highlighting areas of interest – things that Carol and I both had questions about, things that might be significant in identifying him from all the other men called Philip Holmes.

Our thoughts related to these specific areas included:

  1. What significance did the animal have? Was it a working animal? Was it a small horse, a pony or a mule?
  2. Were those possibly tattoos on his arm, or simply odd markings formed from discoloration of the old photo paper?
  3. Philip is well-turned out – his clothes are clean and pressed, he is clean-shaven and well-groomed. What was his purpose in this location.? Working? Vacation?
  4. What is in the background and is it significant?


Philip highlights in photo

Some of these questions we simply mused upon, whilst others led to further investigation. I insisted that Carol send the photograph to someone with expertise in the history of working animals, especially horses. (Google will turn up a suitable expert for just about anything). The expert suggested that this was a mule, and passed our query on to her friend who was a “mule expert”. The mule expert did not have a lot to add to the conversation other than to confirm that this was indeed a mule. He suggested that the background might be indicative of mining in British Columbia around 1920. This was quite exciting, as there was a British Philip Holmes of interest who had traveled through BC and into the US. (After some frantic excitement, including going as far as to contact some of their living relatives, he was ruled out.)

The tattoo theory was interesting.  We were fairly certain at one point that if this was a tattoo, it was possibly a symbol like the caduceus – two snakes entwined around a winged rod, often mistaken or misused as a medical symbol. (The “actual” medical symbol is called the rod of Asclepius after the Greek god associated with medicine and healing. The caduceus is the symbol of Hermes, the messenger god.) Either one seemed like a cool tattoo!

There was nothing to suggest that Philip was involved in manual labor. There appears to be a structure in the background, possibly some kind of tarp shelter or even a campsite. The white objects could be some kind of hard hats like a miner’s helmet, or possibly a pith helmet.

Other than our flurry of excitement over a possible British Columbia connection, all the speculation served little purpose other than to keep the search alive and hopeful.

Meanwhile we diligently checked all of Carol’s mum’s DNA matches across various sites. One day, I was sure, there would be a new DNA match that would help solve the mystery.





Philip Holmes – photographic evidence

To go to the beginning of this story, which is spread over several blog posts, click here.

It’s pretty frustrating to have a pile of family photographs and have no idea who the people are. In the case of Philip Holmes, Carol had an old photograph reputed to be of him. So here we had the opposite – a photograph with a name, but little information on the person.

In some cases old photographs do yield clues  – it can be possible to get an idea of the year/era that a picture was taken by the photographic technique used or the clothes worn by those pictured. Carol had already done some of her own research trying to date the picture.

Philip Holmes reverse of photo Solio crop

The stamp box in the top right is bordered by the word “Solio”. Solio was a photographic paper introduced by Kodak in 1908 and sold all over the world until the 1920s. The photographs were developed directly onto on the photographic paper which was the size and weight of a postcard and had a divided back – one space for a message, the other for the recipient’s address. Philip’s photograph was an example of a “Real Photo Post Card”.

The front of the card is of course where most of the clues lie and we’ll come to that. The back of the card had an oddly worded message which Carol and I discussed at length. It wasn’t going to offer any clues about the identity of Philip but we looked for clues in anything that we could.

Philip Holmes reverse of photo with message

I think we tore apart this message looking for some hidden meaning about the relationship between Philip and his dad. On the one hand, “Dad” was a very informal salutation. It didn’t carry the slight emotional distance of “Father”; nor was it as working class as something like “Da” or as pretentious as “Pater” . The sentiment of “Fondest Love” seemed a little stiff and awkward from the relaxed-looking young man in the photograph. The handwriting looked as if it had been written painstakingly by someone unused to writing messages. And we wondered why the card was written out to “Dad”, not “Mum and Dad”. Had this card, with its very personal family message, ever been given to Philip’s dad?

Philip Holmes part 2. Could it be Phillips Holmes, Hollywood star?

For the first part of this story, follow this link.

When working with DNA, we always say “follow the DNA, not the name.” But when you don’t have any decent DNA matches to work with and you do have a name, it’s extremely tempting to run that name into the ground and come up with theories – some quite wild – about who your mystery relative was.

In her search for her mother’s biological father, Carol had been given a name by her grandmother. This man had been “in the theater”, her grandmother claimed, and his family was quite well-to-do. It turned out that there was someone who fitted the name and some of those details – a Hollywood actor named Phillips Holmes.

Phillips Holmes.

Phillips Holmes was born in 1907 in Grand Rapids, Michigan. His slightly unusual first name was actually his mother’s maiden name. It would seem like a big leap to believe that Phillips Holmes could have been Carol’s grandfather, yet the anecdotal information from her grandmother and a few other biographical pieces of information began to add up.

Carol’s mum was born in London in 1930. This meant that the father had to have been in London before that, for obvious, biological reasons. There was plenty of documentary evidence to show that Phillips Holmes had visited London several times in the years before 1930. Passenger lists also show that he was in the UK after 1931. One shows him returning to New York from Southampton in 1934.

There was, however, no way to prove that Phillips Holmes had been in the UK and met Carol’s grandmother. The photograph of “Philip” in Carol’s possession did not look at lot like the many images of Phillips Holmes on the internet. We even compared their handwriting – Carol’s photograph of her Philip had a message on the back, supposedly Philip writing to his father. We compared it to an image of Phillips Holmes’s signature – they didn’t have a lot in common.

After investigating the parents and ancestors of Phillips Holmes, it didn’t look like Carol, her mum and the actor were connected through DNA either.

There are many ways that we can try to twist details to fit our purpose. Intriguing though it was to find a Hollywood star with the same name as a missing grandfather, it became clear that it was simply a coincidence. There was nothing to connect these two men other than their (almost) identical names.

Philip Holmes – part 1.

Over a year ago I posted a photograph in a blog post about one of my first DNA searches to identify lost family. I promised to write the story of Philip Holmes and the steps we took to identify him – both conventional genetic genealogy steps using DNA and family trees as well as a few more unconventional methods. A few mysteries still remain, including being unable to find any living descendants of Philip – or at least any who will answer our messages on Ancestry. There’s a little more information in this post from last year if you want to read it to get a few more details.

We often remark in DNA searches that you should “follow the DNA” rather than a name. People often have a name to work with. Names can be changed, made up, mis-remembered. DNA cannot be faked or changed. In this case we did have reason to believe that the name of the person we were looking for was indeed Holmes. That’s because it was on a birth certificate.

Screen Shot 2018-05-03 at 10.59.31 AM

Myrena or “MPH” as I have always known to her or even “Carol’s Mum” was born to Ethel Marena French, formerly Ethel Marena Stevens. At the time she had Myrena she was married but had disappeared for a while – possibly a couple of years – and then resurfaced with a new baby in 1931. She always claimed that Mr Philip Holmes was the father of her little girl. The birth certificate reflected the name Holmes although she was not married to him. The baby’s name was also recorded in the birth register under two different surnames – French and also French-Holmes . It’s not often that there is any paperwork connecting the name of an otherwise unknown father to the birth of a child to an unmarried mother. This appears to be a case when it was allowed – or Carol’s grandmother, Ethel, just made sure it happened. Carol did some research and gave me this information:

“The ‘French’ bit was because it was my grandmother’s legal married surname at the time.  Interestingly, there were 2 birth entries; one that says Holmes-French & one that just says French.  I looked into this & apparently it wasn’t that common but was usually when the parents weren’t married but the father was present when the birth was registered & consented to his name being on the birth certificate/entry.”

So it seems in this case there was some further confirmation that we were looking for a man called Holmes. But who was he? DNA at this point had not revealed anything closer than 4th cousins who could be from Carol’s missing maternal grandfather’s side. This did not make it easy to “follow the DNA”. A lot has changed in 18 months of DNA matches on Ancestry. Mainly, though, Brits still face the same issues in that matches are generally US- based and reflect a connection to an ancestor possibly back in the 1700s. So whilst we did do a lot of tree-building and match-connecting, Carol also took the less-favored road:  simply researching men of the same name who would have been around the right age in 1930 to have been her grandfather. She mainly used voter registration lists from London for this, generating a list of around 100 people to work with.

There was reason to believe that Philip Holmes may have spent time abroad. The photographed that sparked my involvement shows a man standing next to a mule in an environment that certainly does not look English. It was this idea of a possible overseas involvement that led Carol to briefly investigate a movie star of the 1930s as her possible grandfather.

Philip Holmes