The final mystery of Richard III: how 21st-century science cracked a 500-year-old cold case

Jo could tell a few things about the skeleton as she, bone by bone, brought it into the light. This appeared to be a male, with injuries incurred at or around the time of death – injuries that could have been sustained in battle. He was aged in his twenties or thirties.

Yet it was when Jo uncovered the spine – and noticed that it was curved – that, as she later told me, the hairs on the back of her neck began to stand on end. Could this skeleton be the remains of Richard III, the monarch so famously described by Shakespeare as a “bunch-back’d toad”? It looked promising. But we had to prove it.

• Quiz | Richard III quiz: how much do you know about the last Plantagenet king?

Our first priority was to lift the skeleton and bring it back to the University of Leicester for more extensive analysis. All the while, the media interest in the project to find Richard III was picking up pace – the press and public were captivated by the possibility of medieval England’s most controversial king residing in a car park. What we were about to announce at a joint press conference by the University of Leicester and the Richard III Society would send this fascination into overdrive.

Yet we couldn’t let the media feeding frenzy distract us. While newspapers waxed lyrical about the sensational discovery of the ‘king in the car park’, we knew we just had to get our heads down and quietly and calmly get on with a rigorous investigation.

Missing for centuries

So how did we go about identifying these remains as Richard III? Well, let’s back up a bit. If you were to go missing, how would your family and friends describe you to a search team? Your loved ones would probably provide them with information such as your last known location, what you were wearing when you left home in the morning, your age, and your hair colour. The Richard III project was essentially doing the same thing – only it was attempting to solve one of history’s most famous missing person cases, and that person had been dead for more than five centuries.

As in a modern case, we had a list of what was known about Richard that would help us identify him. The king’s last known location, as evidenced by historical documents, was in the church of the Greyfriars friary in Leicester, where he was buried after his death in 1485, having been defeated by Henry Tudor at the battle of Bosworth. More specifically, Richard was known to be buried in the choir, a high-status part of the church.

• Read more | Alternate history: what if Richard III had won at Bosworth?

However, more than 50 years later, the friary had been torn down during Henry VIII’s dissolution of the monasteries, and the land had been sold off more than once and built upon. We knew the location of the friary precinct, but the precise whereabouts of Richard’s grave had been lost to history.

Being a king, Richard’s birth and death dates were well known, so we knew he was 32 years old when he died. Shakespeare, in his plays, famously made him out to be a ‘hunchback’ with a limp and a withered arm. But Shakespeare wrote his play Richard III in the early 1590s, more than a hundred years after Richard’s death, so can only have been drawing on previous accounts of the king’s appearance, quite possibly tinged with Tudor propaganda.

• Read more | Have we completely misinterpreted Shakespeare’s Richard III?

Only two of these accounts were from people who, it seems, had actually met Richard III: one of these, John Rous’s renowned Historia Regum Angliae, described Richard as having one shoulder higher than the other. So Richard having a spinal abnormality was a possibility.

This gave us our list of characteristics to look for when identifying Richard. Crucially, we also had to put a number on how likely it was that the skeleton belonged to the king.

We started by sexing and ageing the skeleton. This can be achieved by examining the bones. For example, men’s pelvises are shaped differently to women’s. And while it’s not possible to precisely determine age at death, an examination of the teeth, skull and ends of bones allows an age range to be specified. From this, Jo Appleby and the team determined that the skeleton belonged to a man, aged 30–34 years when he died. The fact that he was male was also supported by the presence of a Y chromosome, which is inherited down the male line.

By examining the bones by eye – as well as using computed tomography (CT) X-ray scanning technology – the team, including forensic engineer Sarah Hainsworth, was also able to detect and examine 11 perimortem injuries (those sustained at or around the time of death). All of these were consistent with being made by weapons from the medieval period. Scans and a reconstruction of the skeleton confirmed the curvature of the spine and suggested that the right shoulder would have been slightly higher than the left.

Something fishy

What we eat and drink affects the make-up of the chemical elements contained within our bones and teeth. The study of this chemical mix is known as stable isotope analysis, and in this case showed that the individual had a high-status diet that had a strong marine (fish) component to it. Meanwhile, radiocarbon dating delivered a 95.4 per cent probability that the subject died between 1456 and 1530, bracketing the date of Richard’s death in 1485.

While Jo and the team studied the individual’s bones, I carried out the genetic analysis of the remains. DNA testing has the potential to unlock the secrets of a skeleton, but there’s absolutely no guarantees: DNA degrades following death, and what remains is often too fragmented and damaged to analyse. Fortunately, with our Greyfriars skeleton, this proved not to be the case.

So how exactly did we use DNA to link this particular skeleton to Richard III? DNA testing for identification purposes in this sort of case involves analysing the DNA from the remains and comparing it with that of a known, usually close, relative. But there was a problem: we didn’t have the remains of any of Richard’s close relatives to hand. The solution was to test against living relatives of the medieval king.

Yet this approach threw up a further complication: Richard himself left no known living descendants. That meant that the DNA we needed would have to come from descendants of other members of his family. And, as mitochondrial DNA is found in the eggs that mothers generate to produce offspring, we required a relative whose relationship to Richard III was by an entirely female line.

Fortunately, we already knew of one such person. His name was Michael Ibsen, and he was descended from one of Richard III’s sisters, Anne of York. Soon our researchers had traced another female-line descendant of Anne, a woman called Wendy Duldig, who also happened to be Michael’s 14th cousin, twice removed. We had our mitochondrial DNA comparators at the ready.

Then came a breakthrough moment: those comparators showed a match with the Greyfriars skeleton. Interestingly, analysis of the Y chromosome, which is passed down through the male line, didn’t show a match to the living relatives we tested.

However, I knew from my years of research in this field, that we might find what’s known as a false paternity, where the biological father is not the recorded father somewhere on the family tree: the genealogy says one thing, but the DNA tells you some- thing else has gone on!

• Read more | Rediscovering Richard III: the story of identifying a lost king

Picking a winner

It was now time to put that all-important statistical number on the likelihood of the skeleton being Richard. And to do that we turned to a method that’s widely deployed in forensic cases such as this: Bayesian statistics.

Bayesian statistics is a way of calculating the probability of something, and updating that probability by bringing in various sources of data or knowledge. An example that’s often used is horse racing. Say you have two horses, A and B, racing one another. If you don’t know anything about either horse, you might conclude that the chance of horse A winning is 50/50.

But what if we could add some more information here by counting how many times each horse has won when they’d been in a race against each other previously? What if you knew that, out of the 12 times these two horses have raced against one another before, horse A has won eight times (67 per cent of the time)? Well, knowing that, you’d say the probability of horse A winning is higher and the chance of horse B winning is only 33 per cent. Each new piece of information you add will change the likelihood of each horse being the winner.

The beauty of the Bayesian method is that it allows us to combine all the evidence together into one calculation to get a statistical number of the likelihood of the skeleton belonging to Richard III. In that, we have only two possibilities: hypothesis 1 is that this is Richard III; the second is that it’s not Richard III.

Remember we had our list of features to look for? Well each one gives us a new bit of evidence we can add to the mix and allows us to calculate what’s known as a likelihood ratio for hypothesis 1 (it’s Richard) versus hypothesis 2 (it’s not Richard).

The first count was: how often do you find skeletons from individuals in high-status parts of churches who are male and aged between 26 and 35? It’s about 18 per cent of the time, if you want to know. We did the same for each category. We looked at the counts of how often you find skeletons with scoliosis in the medieval period, how often you find skeletons with perimortem injuries in high-status parts of churches, and the likelihood of the radiocarbon date (given that we knew the friary was built in the 13th century and closed in 1538).

One piece of evidence that we didn’t throw into the mix was the stable isotope data that showed that this was a person with a high-status diet. We would expect anyone buried in a high-status part of the church to eat better than the vast majority of the population, so adding this to our calculations would, we reasoned, have unduly biased the statistics towards the skeleton being Richard. We also included the fact that the Y chromosome didn’t match.

So, what did it all show? We were lucky enough to have David Balding, a world-famous forensic statistician, on the team calculating the likelihood ratio for us. And the results David produced were, I think it’s fair to say, pretty conclusive. Even with our conservative approach, the likelihood ratio was 6.7 million to 1 in favour of hypothesis 1. That translates to a probability of between 99.999 and 99.99999 per cent that these were indeed the remains of King Richard III. We were pretty excited with the results – and, I would argue, with good reason.

Talk of the nation

We all know what happened next. Media attention intensified further still. Richard III could soon be seen staring back at readers from the pages of some of our most successful tabloids. He was name-checked in Prime Ministers Questions. Rumour has it that he even became a popular topic of conversation in the nation’s pubs! Then, on 26 March 2015, the medieval king was reinterred in Leicester Cathedral in an extraordinary ceremony that was televised around the world.

Yet not everyone was convinced. A handful of people (not one of them, it has to be said, a statistician or geneticist) still didn’t believe we’d found him. Quite often they would pick one bit of evidence (often the DNA) and say it wasn’t robust enough on its own. Or they would claim that the skeleton could have been one of Richard III’s female-line relatives.

• Read more | Richard III: latest DNA study ‘doesn’t prove bones belong to former king’

As part of our analysis, we constructed a family tree, tracing his female-line relatives going back several generations, out sideways and back down to Richard’s generation. By comparing this with historical information, we were able to show that these known relatives were either female (and so could be ruled out), or either too young, too old or too dead to have fought at Bosworth. Or they were recorded as not being at the battle.

So the only person on that tree who could have been at Bosworth was Richard himself. Beyond this, given the rarity of the mitochondrial type, the chances of someone carrying it being alive around the right time, in the right age range, known to be killed in battle, have a spinal abnormality and be buried in the choir of Leicester’s church of the Greyfriars were vanishingly small.

Our critics also seemingly didn’t realise that the beauty – indeed the entire point – of Bayesian analysis is that it allows you to combine all the strands to come up with a likelihood. To put it another way, if you were to go missing, people would describe you by such factors as your age, your height, your build and what you were last wearing. The police wouldn’t go on just one distinguishing characteristic, such as brown hair or blue eyes. Just as we did, they would identify you based on all the strands of evidence together. And the evidence was overwhelming that these were the remains of King Richard III.

Ten years on, I confess I still have to pinch myself that I was even involved in the project. I’m so proud of all that was achieved. The team that carried out the planning, the excavation and the enormous amount of post-excavation work comprised a huge number of academics and non-academics all bringing their own expertise to the table to make the project the tremendous success it was. No one person could have carried out the project on their own. The whole was greater than the sum of its parts as we came together to solve this 500-year-old cold case.

Turi King is director of the Milner Centre for Evolution at the University of Bath. In 2012, she led the verification team during the exhumation and reburial of Richard III.

This article was first published in the April 2025 issue of BBC History Magazine