The Dead Horse Investigation – Update

The Famous Sheboygan Dead Horse Picture

The man, the myth, the legend…over a century later, the questions still linger: Who is the man in the picture, and what is he doing sitting on a dead horse in the middle of the intersection of Indiana and Griffith Aves. in Sheboygan, WI?

Those of you who have read The Dead Horse Investigation:  Forensic Photo Analysis for Everyone, are aware that we were able to narrow down the date of the famous Sheboygan Dead Horse photo to five possibilities:  May 5, 1867, 1872, or 1878, or August 10, 1873 and 1879.  Through a fortuitous encounter with a member of the Saskatchewan Genealogical Society named Susan, we have been able to further narrow the date even further.

A short review for new Dead Horse afficianados…

The intersection of Griffith and Indiana Aves in Sheboygan is quite different now than it was in the late 1800s.  Today there is a traffic circle located just north of where the man was seated on his dead horse.

The area of Sheboygan pictured was the industrial area in town in those days.  Not many records of the development of this area on the south side of the Sheboygan River survive. No earliest date can be obtained from building permits or construction records.  City directories have also been of no use.  They were not published every year, and where they do exist, the listings only reference intersections or streets, without mentioning specific addresses.

Fortunately, there are other clues that can be used to date the picture.  The key to the earliest date the picture could have been taken comes from noticing that the photograph was taken with a wide angle lens. Only using such a lens could the photograph include over half the width of the street, 40 ft.

The wide angle lens was first produced in 1865 in by Emil Busch in Rathnow, Germany.  Of course, it probably took some time for news of the lens to reach Sheboygan, and even more time for the photographer to obtain one. The true earliest date of the picture (the date the photographer received the lens) was probably much later than 1865, but we don’t have enough information to know when this happened. All we can say is that the picture must have been taken after the invention of the wide angle lens in 1865.

The latest date the photo could have been taken comes from comparing the shadows of the buildings to information given by the 1880 census. The census for that year tells us that there were three saloons occupying corners of the intersection.  The northeast corner is occupied by the Evergreen Hotel, the Italianate building in the background to the right that was a hotel and saloon built long before the 1880s.  The southeast corner across from it is empty.

A Missing Shadow

Sheboygan was built on a grid with the streets running north-south and east-west.  The photographer was looking to the north along what was then Griffith Ave, now known as 8th Ave. When a line is drawn from the empty southeast corner directly across the street to the southwest corner, it does not pass through the shadow falling on the man. That shadow is being cast by the building that is on the second lot from the corner. We can’t see what was on the southwest corner, but we can conclude that it was empty.

Since there were at least two empty corners at the time the photograph was taken, it must have been produced before 1880.

The picture was taken between 1865 and 1880.

The time of day can be derived by using the man as a sundial. Measuring the length of the shadow he cast would normally be difficult without a correction for the perspective of the camera.  That is, lengths as viewed by the camera will appear shortened by the sine of the angle of the line-of-sight of the camera with respect to the length being measured.  There would also be a correction for the effects of the wide angle lens, which can also distort lengths.

However as luck has it, the shadows in the picture run directly across the street, so that the sun is exactly in the west (270 degrees azimuth). With the camera facing north-south, the corrections for perspective are much easier.  Actually, if we assume that the photographer was far away from the man and the horse relative to the width of the street, we can neglect the effects of perspective all together. This was probably true, although it might not seem like it because of the way the wide angle lens distorts distances.

To determine the time of day, you also need the height of the sun over the horizon, called its elevation.  This can be determined by using a little bit of high school trig. The line from the top of the man’s hat to the tip of its shadow on the street is the hypotenuse of a triangle.  The line from the same point on the hat to the ground is the side of the triangle opposite the angle we wish to know.  The sine of the angle is the length of the opposite side divided by the length of the hypotenuse.  Taking the arcsine gives us an angle of 15.7 degrees.

You finally need the coordinates of the man and his horse.  According to Google Earth, the man was located about 43 deg 44′ 35″ N latitude, 87 deg 42′ 47″ W longitude.

An ephemeris will tell you that the sun is at 270 degrees azimuth and 15.7 degrees elevation at this location in Sheboygan every year on May 5 and August 10 at 4:52 pm.

To recap, the picture was taken on May 5 or August 10, between 1865 and 1880 at 4:52 pm.

Although the Sheboygan neighborhood shown in the picture was an industrial area, the streets are almost deserted.  To help us narrow down the date even further, we guessed that the picture was taken on a Sunday; otherwise we believe that the street would have been more crowded, if not with people, then with wagons and other horses.

Using the perpetual calendar at, the only instances when May 5 or August 10 fell on a Sunday were May 5, 1867, 1872, and 1878, or August 10, 1873 and 1879.

In the Dead Horse Investigation, we left it at that.  However, during my recent conference in September 2010 in Yorkton, Saskatchewan, a member of the Saskatchewan Genealogical Society named Susan noticed a clue that even the top Dead Horse experts in the world had missed – the leaves on the trees behind the A-frame building in the background to the left.

Because I come from a warm climate, it did not occur to me that the trees in Sheboygan would still be without leaves as late in the year as May.  Since the tree looks like it is in full foliage, the picture must have been taken on either August 10, 1873, or August 10, 1879.

If anyone has any ideas on why the scene favors either one of these dates, please contact me immediately.  In the meantime, buy the book, don’t wait for the movie!

For more information on how you can purchase your very own copy of The Dead Horse Investigation, visit our order desk at

November 26, 2010

Posted in Identifying Old Photographs | Tagged , , , , , , , , , , , , , , | 13 Comments

Bonnie’s New Family

Finding Bonnie’s birth father was one of our most difficult birth parent searches. 

Bonnie and her brother Michael shortly after they arrived in the US in 1957.

Bonnie and her brother Michael knew they were born in Germany in the 1950s, but they did not know they were adopted until 2002.  Although Bonnie had a copy of her birth cerificate, she had not questioned why her parents names did not appear on it, and why it was issued two years after she was born. 

When Bonnie’s mother passed away, a chance remark by a neighbor revealed the truth.  Bonnie explains:

We only found out [we were adopted] because of a neighbor woman, who knew my mother through Bingo.  I was walking my dogs when she approached me and offered her condolences.  She casually mentioned to me that after reading my mother’s obituary in 2002,  ‘I thought your mother said you and your brother were adopted’, because the obituary stated she left a son and daughter.  

If this woman had never mentioned anything, this whole experience would never have unfolded.  That’s what makes our story different from other adoptees.  Most adoptees knew from the get go that they were adopted or at least told by the time they were 18.  Not so with us.  We never knew.

Bonnie’s German birth certificate.

As Bonnie found out when she retrieved her German birth cerificate and her court records, Bonnie and her brother Michael had been adopted in 1957, and brought to the US by their adoptive parents, Alvie and Hildegard Montgomery, a GI and his German wife.  The children were naturalized and grew up in California.  Bonnie discovered that her birth name was Hannelore; her brother’s was Hans Georg. The court papers gave the name of their mother as Martha G___.

Bonnie eventually located Martha living in Tennessee.  Her birth mother had married another American GI, Vern H___, and moved to the United States in the late 1950s, where she had become a naturalized citizen.  Bonnie’s reunion with Martha was friendly.  Martha told Bonnie that after her adoptive mother Hildegard had made arrangements to adopt Bonnie, she found out about Michael.  Hildegard was disraught at separating the sister and brother and begged Martha to allow them to adopt Michael, too.  At first Martha said no- she thought she could raise Michael as a single mom.  But Hildegard persisted and Martha relented.  The two children were adopted together and soon on their way to a new life in America.

"Oliver" could have been Snake Oliver's first or last name. This is only one example of the many records I looked throgh searching for him.

When Bonnie questioned Martha about their birth father, she was not forthcoming.  She gave them few details:  their father was an enlisted man named Snake Oliver from West Virginia.  This would later prove completely untrue.

Bonnie contacted me in April 2009 in response to an article about the amnesiac Benjaman Kyle that appeared in the Indianapolis newspaper the Indy Star.  The article gave my contact information for anyone who might recognize him. 

I knew Bonnie’s case would be difficult.  I began by believing everything Martha told Bonnie although I realized that “Oliver” could be either Bonnie’s dad’s first or last name.  I searched all the miliary records I could access, and contacted the military base in Stuttgart where Bonnie and Michael had been adopted, as well as the VA and American Legion Headquarters in West Virginia.  Yet I found no trace of Snake Oliver. 

Even speaking with Martha’s close friends did not turn up anything.  Martha never mentioned her children; the first time some of her inner circle knew about them was after she died and Bonnie sent a sympathy card.  One of these friends gave me the name of Martha’s niece in Germany with an old address for her.  The niece had since moved.  It took much time to find her current address, but the niece had no further information.  I contacted family members and friends in the US, Germany, and France.  Nearly eveyone who had been close to Martha had passed away; those who were left could offer no help.

In the meantime, Michael took a Y-DNA test, hoping to discover their father’s last name through a match among the hundreds of thousands of entries in the genetic genealogy databases.  But he found no matches that were even close.

Yet I did not give up.   There was still one thing missing from the story.

Marienpflege is the largest orphanage in Germany.

Bonnie had been adopted when she was two years old, but she told me she could not recall where she had been during those two years.  She said she assumed she had been with Martha.  On a hunch one night, I called the Marienpflege, the largest orphanage in Germany.  They could not help me directly, but the person I spoke with gave me the phone number of the Jugensamt (Youth Office) in Stuttgart, where Bonnie and Michael had been adopted.  The Jugenamt in turn gave me the phone number of the Adoptionsvermittlung (Adoption Bureau). 

This was right before the Christmas holidays, so I did not get much further until a few weeks later. When the government offices opened again in January, I spoke to Sybille Breit at the Adoptionsvermittlung who thankfully spoke perfect English.  Sybille informed me that there are three kinds of adoptions in Germany – Catholic, Protestant, and civil, and that she was in touch with the three agencies handling them.  She said she would call them and enquire about Bonnie and Michael’s adoption records.

Like magic, within days, the records appeared in my email box.  Bonnie and Michael were stunned to learn that they had different fathers.  They had assumed their whole lives that they were full siblings, the biological children of the parents who raised them.  It was surprisng enough for them to learn they had been adopted- learning they had different fathers was overwhelming.

Both fathers had unusual names, yet they were not easy to locate.  I sent countless messages to people with those family names on Facebook. I searched LinkedIn and Googled their names.  I contacted the military again with the more accurate information I had.  Bonnie’s father David S___ had been stationed at the Funkerkaserne in Esslingen, but it was closed a long time ago and is now in ruins.  Since neither father appeared in the Social Security Death Index, I knew they were both still alive.

Finally I decided to try the good old fashioned way of finding people and looked in the telephone book.  There were five David S____’s in the United States who fell within the age bracket of her biological father.  On Memorial day I planned to contact them one by one, but did not get very far.  The first David S___ I called was Bonnie’s dad.

Bonnie and her dad.

The next few days were exciting.  David S___ was surprised to hear about his new daughter.  Martha had not told him about Bonnie.  Yet he was as excited about meeting her as Bonnie was about meeting him.  They were in touch at first through email and then over the phone.  A paternity test came back positive making it official – I had found Bonnie’s dad. 

Bonnie, her father, David, and her new sister Susan were reunited on Bonnie’s 55th birthday-a nice day for a happy ending.

Bonnie, her new sister Susan, and Bonnie's daughter Jennifer.

As Bonnie told me:  Thankfully your article about Benjamin Kyle in the INDY Star brought us to you and you were the final tool in getting this whole story unraveled!  Thanks again Colleen for every effort that you’ve made in locating my birth father.

We are still searching for Michael’s birth father, and hope to have a second happy ending in the near future.

Posted in Finding Birth Parents | Tagged , , , , , , , , , , , , | Leave a comment

DNA Rule-Out for Cold Case, Australia, 1970 – Part II

Australian genealogist Deb Cashion

Through a casual exchange of emails a few weeks ago, Deb Cashion alerted me to a recent article “Police chase DNA of Elmer Crawford relatives in Northern Ireland” that appeared in the Herald Sun.  It described the Victoria police search for a DNA family reference for Crawford.    

Since I was traveling in Ireland at the time, I offered to help locate the required reference.  Deb put me in touch with Keith Moor, the Insight Editor of the Herald Sun, an award-winning journalist and the author of the article.  Keith forwarded me two new articles that appeared in the Sun on July 14 “Mystery man was a drifter” and July 15 “Police chase tip to retrieve DNA” that provided more information on the case.       

Elmer about the Time of the Murders

According to the articles, authories were searching for Crawford family members who could serve as DNA references for Elmer.  The object was to compare family DNA to that of an unidentified man who had died in a Texas hospital in 2005 who was thought to be Elmer. 

Under normal circumstances, Y-DNA would have been used to confirm the Texan as Elmer.  Because the Y-chromosome follows the male line of a family as does the family name in western cultures, a Y-DNA match offers the most straightforward method of identifying an unknown male.  Elmer could have aged, he could be living under an assumed name, or he even could have had cosmetic surgery to alter his appearance, but he could not have changed his DNA.        

Y-DNA matching could not be done in the present case, however, because Elmer was illegitimate of unknown paternity, had no known brothers, and his son was dead.  (He killed him).  There were no males who would qualify as a Y-DNA reference, so that mitochondrial DNA (mtDNA) would have to be used instead.  Because this type of DNA is handed down along the direct female line of a family, it can be difficult to locate a family member who can serve as an mtDNA reference because the name changes in each generation.  

In my experience, the key to success in soliciting someone to give a DNA sample is being well prepared with information about his family pedigree. My first step was to learn as much about Elmer himself as I could. 

Elmer Kyle Crawford Baptismal Record

Elmer was born 18 May 1929 in Quebec, Canada, the son of Anna B. Crawford and an unknown father. His mother traveled to Canada from Tamneymartin, Maghera Parish, Co. Derry, Ireland, to give birth to Elmer.  This information allowed me to locate his baptismal record in the Drouin Collection, the largest repository of French Canadian church records.  He was baptized 11 June 1929 in Hemmingford, Quebec.  One of his baptismal sponsors was his maternal grandmother Elizabeth Crawford. The Sarah Crawford whose burial record appears on the page following Elmer’s baptismal record was probably a relative, perhaps the reason Elmer was born in Hemmingford.

During a brainstorm, I emailed the Maghera District Genealogy and History Society for “Help with a Murder Case”.  An immediate response came from Society member Denver Boyd, “We don’t normally get emails as dramatic as yours!”

Denver provided a pedigree of the Crawford family.  It seemed ordinary even though one of its branches ended with a cold-blooded killer.  He also provided me with the contact information of Gilbert Crawford, a local businessman, auctioneer, and Justice of the Peace.  He is also Anna B. Crawford’s nephew.  If anyone knew everyone, it would be Gilbert.

In the next few days, with the help of Denver and Gilbert, I discovered that Anna’s sisters had not married and had had no children.  It was time to research one generation back, with the hope of finding a female line that had survived. 

Descendants of Robert Kyle, Elmer Crawford's Great Grandfather

Elizabeth Kyle Crawford’s family was large.  She had four sisters and three brothers.  It seemed there was ample probability of success.  Yet according to the family chart, of those four sisters, the only one to marry besides Elizabeth was her younger sister Charlotte.  Charlote had had three sons and one daughter Kathleen, but Kathleen was deceased and had had no children.  I was back to square one.  

Maghera Parish, Ireland

In a foreign country, internet access can be difficult when you are staying in B&B’s as we were.  Libraries usually have terminals for their patrons to check email and surf the web, but it is usually not possible to connect your laptop to their server.  If you have to send email attachments, libraries are not much use to you.  Drogheda, the largest town near the village where were were staying, did not have an internet cafe, but we were lucky to discover a hotel where I could purchase broadband access.  Since my cell phone did not work in Ireland, my personal calls to the Crawford family had to be made over Skype from the hotel.  This was a quiet and comfortable arrangement, but it did not allow anyone to call me back.  

Nevertheless a lucky mistake led to the sought-after mtDNA reference.  After speaking to Gilbert’s sister June, the Crawford family genealogist, I discovered that I had forgotten to ask her for a certain phone number.  I was glad I called her back.  June had just located her genealogical documents, and discovered that Elizabeth’s sister Sarah had not died in 1901, as indicated by the family chart.  The notation shown in the chart was only an abbreviation.  The attached pedigree that I had ignored revealed that Sarah had died after 1901.  June discovered that Sarah had married, had had two daughters and a son, had lived until April 1979 and that Sarah’s granddaughter Lois still lives in Maghera Parish.  By coincidence Lois’ daughter Sheelagh lives in New South Wales, Australia making it convenient for the Victoria police to contact her for a DNA sample.  

Although it sounds as if everything fell into place at one time, it took me several days to work out the family connections.  We were traveling, and the hotel internet proved too expensive for everyday use.  Luckily we discovered that the local McDonald’s offered free internet access with the purchase of a cup of coffee.  Free parking was included.  Over the next few days, we drank a lot of McDonald’s coffee as I emailed to Keith Moor and the Victoria police Sheelagh’s contact information and the documentation confirming her as Elmer Kyle Crawford’s mtDNA reference.  My job was done.

Elmer Kyle Crawford, c 1970

It took a few weeks for the authorities to obtain a DNA sample from Sheelagh, analyze it, and discover it did not match the man who died in the Texas hospital.  The man in the hospital is still unidentified, and Elmer Kyle Crawford, who so brutally murdered his young family in 1970, is still at large.   

If you have any information on the whereabouts of Elmer Kyle Crawford, please contact me at  You may remain anonymous.  

DNA Rule-Out for Cold Case, Australia, 1970 – Part I 


Posted in Crime and Law Enforcement | Tagged , , , , , , , , , , , , , , , , , , | 8 Comments

DNA Rule-Out for Cold Case, Australia, 1970 – Part I

James, Theresa, Karen, and Kathryn about the time of the Loch Ard Gorge blowhole, Port Campbell Park, Victoria, Australia

A crime that could not escape media attention in Australia, any more than the Manson family murders could have evaded the public eye here in the States-Elmer Crawford brutally murdered his pregnant wife Theresa (35) and his three children Kathryn (13), James (8), and Karen (6) in 1970. He then disappeared.   

He electrocuted them while they slept using an electrical cable he had fashioned from an extension cord and alligator clips he attached to various places on their bodies, then bashed them with a hammer.  He then drove nearly four hundred miles with their bodies in the back of his car, and rolled it over the side of the Loch Ard Gorge in Port Campbell National Park, Victoria.  He was probably hoping the car would vanish into the Gorge’s spectacular blowhole where the sea surges underground for 100 metres and then erupts through a huge hole, 17 metres deep and 40 metres wide.  To his misfortune, a drainage ditch on the cliff overlooking the Gorge prevented Elmer from easily rolling the car off the edge of the cliff.  The crude ramp he built to maneuver it over the ditch caused the car to land on the embankment below, thwarting what might have otherwise been a perfect crime.    

Loch Ard Gorge blowhole, Port Campbell Park, Victoria, Australia

 Elmer’s motive for committing such an horrific crime may never be known, but an unfinished letter written by his wife indicated his unhappiness with her pregnancy, and that she had recently discovered that Elmer had been stealing from the Victoria Racing Club where he had been employed for years.         

Fast forward to 2008, when Australian genealogy sleuth Deb Cashion noticed a resemblance between an age-progressed picture of Elmer in a Victoria’s Herald Sun newspaper article and that of an unidentified man who had died in a Texas hospital three years earlier. Facial recognition experts with the Victoria police and the FBI in Quantico, VA agreed that there was a close match beween the two, but sought a DNA test to confirm the identification.  

As an experienced photo analyst, I am aware that facial recognition is not an exact science, nor are age progression techniques.  Two people can strongly resemble each other, yet not be related at all.  How many times have you run into someone who looks enough like you to be your twin?  Age progression techniques have been used to produce highly accurate renditions of what individuals look like at an advanced age, based on the aging patterns of their parents and their psychological profiles.  But until a person is found, is it not possible to know how well his appearance as an older person has been predicted.  Yet facial recognition and age progression techniques can be used to rule-out that two people are the same, by comparing features such as the shape of the hairline and eyebrows, and the width of the nose bridge.  The folds in earlobes can also be useful.  If selected features mismatch, the two people cannot be the same.  

Elmer about the time of the murders, what he might look like today, and the man who died unidentified in a Texas hospital

The unidentified Texan could not be ruled out as Elmer Crawford.  Both men were the same height, had the same eye color, and tilted their head at the same angle.  They also had the same deformity of their left ears.  The Texan’s hairline was similar to that of the age-progessed photo of Elmer, as were his wrinkled forehead and jowls.       

Yet even a close match between the two did not prove they were the same person.  Since the man in Texas had obliterated his fingerprints, fingerprinting could be not used.  A DNA test was needed.  That’s where I came in. 

To be continued…   

DNA Rule-Out for Cold Case, Australia, 1970 – Part II

Read most recent Herald Sun Articles:  

DNA tests confirm unidentified body in Texas is not Elmer Crawford          August 26, 2010  

On the trail of killer dad Elmer Crawford                                                                   August 26, 2010 


Posted in Crime and Law Enforcement | Tagged , , , , , , , , , , , , , , , , , , | 2 Comments

Unknown Child on the Titanic – Part IV (Conclusion)

1901 Census Frederick and Augusta Goodwin Middlesex, Edmonton

When AFDIL attempted an identification through Y-DNA, I was asked by my colleague Dr. Odile Loreille to find a Y-DNA reference for Sidney Goodwin.  We were just finishing up the identification of The Hand in the Snow, so she knew I was available for a new project. Of course, my first step was to search to obtain information about the Goodwin genealogy.  I immediately found Sidney’s parents Frederick and Augusta in 1901 living in Middlesex with their four oldest children Lillian (5), Charles (4), William (2), and Jessie (1).  Frederick was listed as a print compositer. 

1891 Census Charles Goodwin and Family, London. His son Thomas had alread left for America.

Because Frederick and his sons perished on the Titanic, to find a Y-reference for the family I researched Frederick’s three brothers to see if they had any living male descendants.  Frederick’s older brother Thomas (b. 1869), emigrated to Niagara Falls, NY in the mid 1880’s.  It was on Thomas’ suggestion that Frederick had decided to move with his family to the United States to work at a new power station opening in 1912.  Unfortunately, Thomas only had two daughters. 

Frederick had two other brothers, Sidney, who was mentally retarded, and Frank, who never married. 

Since there were no living males descending from Frederick’s brothers, I researched Frederick’s paternal uncles for surviving male lines.  Frederick’s father Charles had several brothers, including Samuel, b. 1838. 

1861 Census Thomas Goodwin and Family. His son Samuel (b 1838) had already left home

I was surprised to find a website dedicated to Samuel Goodwin and his family. The site had been posted by a genealogist named Mary from the Netherlands, who had taken an interest in the Goodwin family story.  Mary was kind enough to put me in contact with Carol Goodwin, the matriarch of the Goodwin family, whose grandmother had been the sister of Frederick Goodwin.  Carol had been working on a book Titanic’s Unknown Child about the Titanic Goodwins, and provided me with much valuable genealogical information about the family that was useful for tracing a family reference. 

Carol Goodwin, Author of Titanic's Unknown Child

Mary told me that about ten years ago, she had been in touch with a Graeme Goodwin in Australia who was the grandson or the great grandson of Samuel Goodwin of the Titanic family, but she had lost contact with him after he changed his email address.  Fortunately, she was able to remember his middle initial and that he lived in Queensland with his sister.  This provided me with enough information to locate him.  There were only two G. A. Goodwins in the Queensland telephone directory, and he was the second one I called.  He confirmed that he was the Graeme Goodwin whose family members had been lost in the Titanic and who had been in touch with Mary years earlier.  Graeme was delighted to be back in contact with his extended family.

Photo of Samuel Goodwin, his wife and family from Christchurch, NZ

Graeme had a photograph of Samuel Goodwin and his family that was published in the newspaper in 1908 on the occasion of Samuel’s 50th wedding anniversary.  The caption said the family lived in Christchurch, NZ.  Graeme told me that Samuel had originally immigrated to New Zealand, but that his own branch of the family had moved to Australia more recently.  Samuel’s picture helped me locate a second Goodwin living in Dunedin, NZ as a backup.  Because there is a chance of an unrecorded adoption, name change, or illegitimacy in a family, it’s good to have two DNA references for an identification.

Meanwhile back at AFDIL, Dr. Rebecca Just, Dr. Odile Loreille and their team of researchers made an attempt to distinguish the two children using the mtDNA coding region, since analysis of the HVR1 and HVR2 control regions had failed. They used two commercially available SNP assays that had proven useful in differentiating HV haplotypes.  Yet the coding region results of the Panula and Goodwin families references were identical.  

As a last hope, the team used a more targeted approach, studying 92 published mtDNA genomes with the HV haplotype, searching for regions that had high levels of inter-individual variation. They discovered a region bounded by positions 8,164 and 11,160 that had not been covered by the two standard SNP assays.  In this region, they discovered a rare mutation at position 9923 that the Goodwin reference had, but that the Panula reference lacked. 

Sidney Leslie Goodwin, the Unknown Child on the Titanic

The remains also exhibited this rare mutation.  The tie was finally broken.  The Goodwin reference matched the remains in HVR1, HVR2, and in the coding region.  The Panula reference had one mismatch in HVR2, and a second mismatch in the coding region at position 9923, giving the two differences required by forensic guidelines for a rule-out.  After 90 years, the Unknown Child on the Titanic finally had a name.  He was Sidney Leslie Goodwin.


In August 2008, led by matriarch Carol Goodwin, the Goodwin family met in Halifax for a memorial service to honor Sidney and the other 52 children who died on the Titanic.  Goodwin family members came from California, Wisconsin, New York, and England to attend the ceremony.  I was invited to attend as I had been made an honorary member of the family.  

Sidney’s memorial service was held at St. George’s Anglican Church, where the Unknown Child’s funeral had taken place in 1912.  The family assembled at Sidney’s grave in Fairview Lawn Cemetery, where the service was continued. A bell was rung as the name of  each child who died was read by members of the family. 

The Goodwins at Sidney's Memorial, August 2008

As an honorary Goodwin, I was privileged to hold Sidney's shoes.

Our Goodwins also visited the Martime Museum of the Atlantic, where one by one each of us was allowed to hold the tiny shoes that had provided the key to reversing what could have been a serious historical error that would have gone tragically uncorrected. 


In 1911-1912, a pair of small brown shoes was manufactured somewhere in England. It made its way to a retail shop where Augusta Goodwin bought them for her baby son Sidney to wear on the family’s journey across the Atlantic where her husband Frederick had the promise of a new job in America. 

Sidney would die wearing those shoes a few short months later, along with Augusta, Frederick, and their other five children:  Lillian (16), Charles (14), William (11), Jessie (10), and Harold (9) , victims of the worst maritime disaster in history.  They might not have perished if the politics of the times had been different.  The family had booked third class passage on a small steamer out of Southampton, but due to the coal strike that year, the voyage was cancelled and the family was transferred to the Titanic.

Ninety years later, the ordinary little shoes that Sidney wore the night he died would become an extraordinary key to one of the most compelling stories of human identification.

Yet even more important than his shoes, was the blueprint Sidney carried in each cell of his body that defined who he was-the blueprint called DNA.  It would take decades before DNA identification would be discovered, even as Sidney’s remains were dissolving into the soil where he was laid to rest in 1912. It would take even more years for DNA analysis to mature into a sophisticated science that had a chance of identifying Sidney’s remains, and by that time only crowns of three of his milk teeth and a small bone shard would be left.  And even as the small amount of DNA that could be extracted from one of those crowns and from the bone was consumed in multiple rounds of testing, it would take the stubborn persistance of ancient DNA specialists and scores of genealogists from around the world to finally identify him based on only one picogram of DNA – about the amount of DNA present in only a single cell of his body. 

And now, even that is gone.

Part I, Part II, Part III, Part IV (Conclusion)

Posted in Historical Identifications | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | 15 Comments

Unknown Child on the Titanic – Part III

A Cell with Nuclear and Mitochondrial DNA

To understand what happened next, you have to know a little about mitochondrial DNA (mtDNA).  Mitochondrial DNA is contained in small, football-shaped inclusions outside the nucleus of a cell. It’s widely believed that mitochondria were once independent bacteria that invaded primitive cells millions of years ago.  Instead of being digested, these bacteria took up residence in the cell, forming a symbiotic relationship with it.  The cell provided them with food and water, and the mitochondria provided the cell with energy for metabolism and heat.  The arrangement worked out so well that millennia later, a human cell has up to 1,000 mitochondria, each carrying five to ten copies of its own genome.    

A Mitochondrion

Mitochondrial DNA is passed along the exclusively female line of the family.  A mother passes it to all of her children, but only her daughters pass it on to the next generation.  My brothers and my sister and I have my mother’s mtDNA, but only my sister has passed it on to her children.  My brothers’ children inherited their mtDNA from their mothers, who are my sisters-in-law.  

Not many people understand why this happens.  An egg cell is large, containing hundreds of mitochondria.  The sperm is small, with only a few in its tail.  In the act of fertilization, the sperm injects its genetic material into the egg, but is otherwise destroyed along with its mitochondria.  The fertilized egg divides, becoming an embryo that develops into a fetus that eventually is born as a child carrying the mitochondrial DNA of only his or her mother.    

Back to the Unknown Child…    

The mitochondrial genome is shaped as a ring, 16, 563 or so bitpairs (bp) long.  The positions are numbered starting at the “top” with 1 clockwise to 16,563.  Not everyone has exactly 16,543 bit pairs in their mtDNA genome.  It is common for the mtDNA genome to experience insertions and deletions, making it slightly longer or slightly shorter than this.    

Mitochondrial DNA is abundant compared to Y-DNA, the other type of DNA often used for identifying males.  There can be up to 10,000 copies of the small (16,563 bp) mtDNA genome per cell, but there is only one Y-chromosome, about 50 million bp in length.  Identification using ancient or degraded DNA usually relies on mtDNA, since the probability of obtaining enough of the right kind of Y-DNA for analysis is usually much lower. 

The two sections of mtDNA that are used for identification purposes are known as Hypervariable Regions 1 and 2 (HVR1 and HVR2).  These make up the “control region” of mtDNA.  They are well-characterized and easy to work with.  HVR1 is the segment of mtDNA extending between positions 16001 and 16563.  It is usually the segment tested first. 

When the original mtDNA tests had been performed on the Unknown Child, only the HVR1 region had been tested.  That is, only low resolution tests had been done.   This had been sufficient to rule out four of the children.  However, because the DNA profile (called the haplotype) of the Unknown Child is very common – the HV haplotype is shared by about 15% of Western Europeans – it is not surprising that the HVR1 results of two out of the six candidates matched those of the Unknown Child.  Further DNA testing should have been done before an identification was announced, but because the age estimate given by the teeth was thought to be reliable, it had not been considered necessary.  

After doubts arose about the identity of the child, a new round of DNA testing was performed on the HVR2 region, extending between positions 1 and 574.  It is usually the second region that is tested to improve the resolution of the analysis.  The results of the second round of testing again ruled out the same four children.  They also showed that the HVR2 profile of the Goodwin family reference matched that of the remains, but the Panula reference showed one difference, at position 146.  While this might seem to exclude Eino Panula, forensic identification guidelines suggest that an exclusion be made on the basis of two differences because of the high mutation rate of mtDNA.  Family references are often far removed in the famly tree from the person whose remains may be under examination, so that it is possible that in the intervening generations a random mutation can occur in the family genome. Ruling out the Panula baby based on a single mismatch could result in another mistaken identification, considering how common the haplotype was, and the contamination that had been present in the earlier round of analysis. 

So in forensic terms, the second round of testing was still tied.

There were only two remaining DNA tests possible.  One was a last attempt to obtain Y-DNA from the remains, but there was only a remote chance that any Y-DNA had survived, and there had never been any effort made to locate paternal references. The focus had been on finding mtDNA references for the candidate children.  The second hope was to test the coding region of mtDNA, the region outside the HVR1 and HVR2 control region.  However, the first two attempts to distinguish the children using standard testing of the coding region failed.  We were quickly running out of options.

To be concluded…

Part I, Part II, Part III, Part IV (Conclusion)

Posted in Historical Identifications | Tagged , , , , , , , , , , , , , , , , , , , , , , , , | 4 Comments

Unknown Child on the Titanic – Part II

After eight and a half decades, there was little left of the child’s body. Only a small piece of wrist bone and the crowns of three tiny baby teeth had survived the inclement weather and damp, slightly acidic soil.

In the spring of 2002, when Parr and Ruffman determined that the child was not Gosta Paulson based on a mismatch between the mitochondrial DNA (mtDNA) obtained from the bone shard and DNA provided by a maternally-linked Paulson relative, the teeth became more significant in the identification efforts.  Dr. E. J. Molto, an anthropologist and the director of the Paleo-DNA Laboratory at Lakehead University, suggested that the three teeth belonged to “quite a young child”.  The teeth were sent to Dr. Christy Turner at the State University of Arizona in Tempe, who agreed with his assessment.

Alfred Edward Peacock

There were five other male children about two years old or younger who died on the Titanic:  Eugene Francis Rice (2 1/2 years), Sidney Leslie Goodwin (19 mos), Eino Viljami Panula (13 mos), Alfred Edward Peacock (7 mos), and Gilber Sigvard Danbom (5 mos). Parr and Ruffman concentrated on finding maternally-linked relatives of the two youngest children for DNA comparison.  When in the summer of 2002, these two children were ruled out by mtDNA analysis, it was time to take a much closer look at the teeth.

Bruce Pynn, an oral and maxillofacial surgeon in Thunder Bay, suspected that one of the teeth might contain dentin from which additional mtDNA could be extracted.  Having a second mtDNA extraction would confirm the DNA profile obtained from the first bone-shard extraction, which had been partially contaminated.  Pediatric dentists Keith Titley and Gajanan Kulkarni, and dental anthropologist John Mayhall at the University of Toronto identified the three teeth as the maxillary right second primary molar, the mandibular left primary cuspid, and the mandibular right first primary molar of the child.  Furthermore, based on the state of development of the crowns, the lack of root development, and the absence of wear, the teeth were estimated to have come from a child between 9 and 15 months old.   Eugene Rice was ruled out not only by an mtDNA mismatch, but also because he was 2 1/2 years old.

Tooth #3 from below

Tooth #3 from below.

Upon examination by a scanning electron microscope, one of the teeth revealed dentin where the enamal layer had flaked off.  The dentin was visible around the edges of the interior of the tooth, with debris filling the pulp chamber. The tooth was sent to Dr. Scott Woodward at the ancient DNA laboratory at Brigham Young University in Provo, UT, where a new mtDNA extraction was performed.  The mtDNA obtained matched the original mtDNA that had been extracted earlier from the bone shard, confirming that the earlier analysis had been correct.

Eino Viljami Panula

Two children remained, the 13-month-old Finnish child Eino Panula, and the 19-month-old English child Sidney Leslie Goodwin.  The mtDNA of the family references for both children matched that obtained from the remains, indicating that they had a common maternal ancestor within the past 2,000 years.  Neither child could be ruled out on the basis of the mtDNA data alone, but taken together with the evidence of the age of the child provided by the teeth, on November 6, 2002, Parr and Ruffman announced that the Unknown Child on the Titanic had been identified as 13-month-old Eino Panula.

Doubts soon arose about whether the identification was correct, however, based on the shoes of the child that had been preserved at the Maritime Museum of the Atlantic.  The shoes were thought to be too large for a 13-month-old child; they would have certainly fallen off before the body was recovered almost a week after the disaster.

The shoes of the Unknown Child

Clarence Northover, a Halifax Police Department Sergeant in 1912, helped guard the bodies and belongings of the Titanic victims. According to his grandson Earle Northover, “Clothing was burned to stop souvenir hunters, but he was too emotional when he saw the little pair of brown, leather shoes about fourteen centimeters long, and didn’t have the heart to burn them. When no relatives came to claim the shoes, he placed them in his desk drawer at the police station and there they remained for the next six years, until he retired in 1918.”  Clarence Northover moved to Ontario when he retired as Deputy Chief in 1919. In 2002, his grandson Earle Northover decided the shoes belonged back in Halifax and donated them to the Museum, which performed extensive research to authenticate the shoes before accepting them into its collection.

The 1912 Coroner’s report of the child included a description of a pair of “brown shoes”.  The identification of the shoes as those of the Unknown Child was supported by research through catalogues and consultation with clothing and footwear museums to show that the style of the shoes are appropriate for the period, roughly 1900 – 1925, and that they were likely manufactured in England.  Chemical tests were made to look for traces of seawater and an electron scanning microscope was used to search for saltwater diatoms but the results were inconclusive. The testing found large amounts of salt on the shoes, but the trace elements did not exactly match the proportions in sea water. The testing lab suggested that the chemical components may have been distorted by salts in the tanned leather, by washing or by abrasion.

According to Dan Conlin, the Museum’s Curator,

“We hear from people all the time who think they have objects from Titanic.  Unsually nice people who have something in the family or bought something from an antique shop which they think or hope is from Titanic:  key tags, steward’s badges, door plates, bells, belt buckles and more deck chairs than I can remember.  They inevitably turn out to be wishful thinking.

“However, Northover’s shoes seemed very different from the start.  The family’s story had the ring of plausibility connected to a significant person and institution in Halifax of 1912.  Earle’s grandfather had retired  as Halifax’s deputy polie chief.  He told his sons, who told their children how grandfather Clarence had guarded the Halifax morgue where Titanic’s victims were brought after the sinking.  When the discarded clothes were swept up, he did not have the heart go burn the tiny pair of baby’s shoes but kept them in a box at his desk in the police station and they went to Ontario with him when he retired.

“Museums have a high standard of authenticity  Starting with the Northover’s oral history, we searched newspapers, city directories, police personnel records, coroner’s reports, period footwear catalogues, shoe historians on two continents and the latest in scientific testing.  Taken together, the documentary evidence confirmed Northover’s story.”

If the shoes had indeed belonged to the Unknown Child, could the identification of the child as Eino Panula have been in error?  While the assessment of the age of the child based on the teeth was the opinion of a group of world odontological experts, it was still subjective.  The initial DNA analysis, although objective, had indicated that the child could be either the Panula or Goodwin baby.

The responsibility of identifying the child now returned to the ancient DNA community, with the hope that additional DNA analysis could genetically differentiate between the Panula and Goodwin families.  Identification of the tiny baby who died decades ago now depended on some of the most sophisticated technology on earth.

To be continued…

Part I, Part II, Part III, Part IV (Conclusion)

Posted in Historical Identifications | Tagged , , , , , , , , , , , , , , , , , , , , , , , , | 7 Comments