Science Vs - DNA and the Smell of Death
Episode Date: October 21, 2016In these cases, emerging DNA evidence and the smell of death (yes, really) pushed the boundaries of what was technologically possible. But how reliable are they? To find out, we go to a body farm and ...talk to Assoc. Prof. Joan Bytheway, Asst. Prof. Sheree Hughes-Stamm, Matt Young, Dr. Arpad Vass, and Asst. Prof. Donovan Haines. Our Sponsors Lenovo – See how Lenovo is revolutionizing data center technology. Modcloth - Enter promo code SCIENCEVS at checkout to get $20 off an order of $100 or more! Squarespace – The easiest way to create a beautiful website, portfolio or online store. Use the offer code “SCIENCE VS” to get 10% off your first purchase. Wealthsimple – Investing made easy. Get your first $10,000 managed for free. Credits This episode has been produced by Wendy Zukerman, Shruti Ravindran, Diane Wu,and Heather Rogers. Our senior producer is Kaitlyn Sawrey. Edited by Annie-Rose Strasser and Caitlin Kenney. Fact checking by Michelle Harris. Thanks to Joseph Lavelle Wilson, Will Doolan and Beth McMullen. Sound design and music production by Matthew Boll, mixed by Martin Peralta and Matthew Boll Music written by Bobby Lord. Selected References 2009 National Academy of Sciences and 2016 President’s Council of Advisors on Science and Technology reports on forensic science. How DNA is transferred in trace evidence. Report on error rates in DNA forensic analysis. Study that showed dogs could pick up the smell of a corpse 667 days later. Call to arms on improving forensic science: editorial. Scent of death - Belgian paper that found three out of four of Dr Arpad Vass' “human specific markers” in other animals. Learn more about your ad choices. Visit podcastchoices.com/adchoices
Transcript
Discussion (0)
After decades of shaky hands caused by debilitating tremors,
Sunnybrook was the only hospital in Canada who could provide Andy with something special.
Three neurosurgeons, two scientists, one movement disorders coordinator,
58 answered questions, two focused ultrasound procedures,
one specially developed helmet, thousands of high-intensity focused ultrasound waves,
zero incisions, and that very same day, two steady hands. From innovation to
action, Sunnybrook is special. Learn more at sunnybrook.ca slash special. What does the AI
revolution mean for jobs, for getting things done? Who are the people creating this technology?
And what do they think? I'm Rana El-Khelyoubi, an AI scientist, entrepreneur, investor, and now host of the new
podcast, Pioneers of AI. Think of it as your guide for all things AI, with the most human
issues at the center. Join me every Wednesday for Pioneers of AI. And don't forget to subscribe
wherever you tune in. It's season three of The Joy of Why, and I still have a lot of questions.
Like, what is this thing we call time?
Why does altruism exist?
And where is Jan Eleven?
I'm here, astrophysicist and co-host, ready for anything.
That's right.
I'm bringing in the A-team.
So brace yourselves.
Get ready to learn.
I'm Jan Eleven.
I'm Steve Levin.
I'm Steve Strogatz.
And this is... Quantum Magazine's podcast,
The Joy of Why.
New episodes drop
every other Thursday,
starting February 1st.
Hi, I'm Wendy Zuckerman,
and you're listening to
Gimlet Media's Science Versus.
This is the show where we pit facts
against law and order.
Quick warning, we're going to get our hands on some dead bodies
in this episode.
So if you have some small ears listening or some big ears
on a small person, maybe you want to hold off listening
to this episode right now or switch to headphones.
This is part two of
our episode on forensic science. In part one, we looked at areas of forensic science that have been
used for decades. These include bite marks, fingerprints, and microscopic hair analysis.
So if you haven't heard that one, we recommend that you go back and have a listen. Now in this
episode, we're looking at forensic science that has just started entering the courtroom and putting people behind bars.
Forensic science that is pushing the boundaries of what's technologically possible.
To explore these new emerging sciences, we're going to dive into two high-profile cases that you might have heard of.
First, we'll look at the Amanda Knox trial to explore the emerging problems with DNA evidence.
And then we'll look at the Casey Anthony trial
to investigate cutting-edge science into the smell of death.
Yep, you heard me right.
The smell of death.
When it comes to whodunit, everyone has an opinion, but then there's science.
My producer Caitlin Sori and I headed to Texas to meet some researchers that are working at the
forefront of forensic science. Welcome to Texas. We're in Huntsville, about an hour's drive north of Houston,
and dotted in the field surrounding us are cows as well as seven prisons,
including one that houses inmates on death row.
We're driving a little bit away from the main town,
heading to a research facility run by Sam Houston State University.
It's called a body farm,
and it's where donated dead bodies are carefully laid on the ground outside so that scientists can watch as the corpses break down.
Caitlin was a little nervous. It's like we've just driven into a horror
film. Come on! No but it is, like look at all these cute little
houses, there's mist in the air and we're on a
road towards a body farm.
It's a bit creepy. You are absolutely
reading into this. This is just a small town.
We're driving to our deaths.
Dead end.
We are literally seeing a sign that says
dead end. You might want to slow down there.
Yeah.
The researchers at the body
farm study the corpse's smell,
skin, hair, nails and bones.
Their findings are then used to help figure out
what cops can tell from dead bodies at a crime scene.
So what can we do for you?
The head of this research facility is Dr Joan Bytheway
and we met her in a room full of bones,
carefully put away in little drawers.
Doesn't it smell good?
What is that smell?
Just from the bones and, yeah, that's what it smells like.
It's a wonderful space in here.
Joan wore bright pink scrubs.
Her glasses' chains were sparkling with little beads.
Joan was training to become a hairdresser before she completely switched gears,
eventually doing a PhD in anthropology and ending up here.
Joan took us on a tour of where the bodies are being left to rot.
Hayley Dodd and Kevin Durr, who work with her,
came along for the ride.
My rule that I always have when I come out here
is I always bring gloves, always,
because you never know. You never know.
This facility is surrounded by a high metal fence with razor wire running along the top,
kind of like how you see at prisons.
The vultures have heard us approaching
and cleared off for the moment.
Through the gate is a forest of tall trees and some shrubs.
Caitlin and I are playing it cool
until we spot our first
corpse. So is this person
wearing a shirt? No, that is
what the flesh ends up looking like throughout the
decomposition process. It looks like they're wearing
clothes, but it's actually skin.
It's desiccated skin, so it's
lost all of its moisture
and it's incredibly tough when it's in that
state. Incredibly tough.
And that skin can last like that for years,
when you've counted it out here.
Most of the bodies we're looking at today
have been decomposing for at least a couple of months.
They are stretched out like starfish.
Joan tells us that you could find a lot of evidence
on a body like this if it was at a crime scene,
including, say, the DNA of a
perpetrator. Let's say that this individual had physically tried to defend herself or and she had
DNA under her nails. You know, you could potentially, because the nails are very resilient,
I mean, they will stay on the hands and feet and you could potentially use that to collect DNA.
But when you scrape under the fingernails of a victim
looking for a perpetrator's DNA,
there tends to be only a small amount of DNA to work with.
Over the last decade or so,
scientists have become much better at analysing that small amount of DNA,
which, by the way, is called trace DNA.
And it can be found in the handful of cells that you shed when you sweat, cough or even
just touch a doorknob.
And trace DNA is the first type of forensic science that we're diving into today.
We're asking, how reliable is the evidence that we get from trace DNA? To explore this, let me tell you
about what happened on November 2007 in a small, picturesque Italian town. A British student,
Meredith Kircher, was found stabbed to death in her house. The alleged murder weapon was
found in the kitchen of a young man called Raphael Solicito.
Raphael was dating one of Meredith's housemates, Amanda Knox. An Italian investigative team said
that trace DNA on the blade of the knife matched the victim, Meredith, and they said that DNA on
the handle of the knife matched Amanda. These DNA traces
painted quite the picture in people's minds. Amanda Knox holding a knife and plunging it
into her housemate. A bloodied bra clasp was also uncovered. It was Meredith's,
and the investigators said it had Raphael's DNA on it. With these key pieces of evidence,
the Italian police charged Amanda Knox and Raphael Solicito
with Meredith's murder.
The story had the hallmarks of a juicy one.
Hot international students, young love and murder.
And the media lapped it up.
American student Amanda Knox is found guilty of killing a fellow student
during a drug-fuelled sex game that went horribly wrong.
The girl known as Foxy Noxy is why Italians are calling this the trial of the decade.
Amanda and Raphael were found guilty of murder.
But then, after spending four years in prison,
Amanda Knox and Raphael Selecito were acquitted,
partly based on problems with the DNA evidence.
Just last year, Italy's high court annulled the convictions.
So, what went wrong with the DNA evidence?
To find that out, we have to start with some forensic DNA 101.
How do you match tiny bits of DNA, trace DNA,
found on a knife or a bra, to an individual?
Our guide here is Dr Cherie Hustam.
She works with the bodies at the Body Farm,
but we met her at a conference room at the university.
So we came all the way to Texas.
Hi, how you doing?
Great, how are you?
Good.
Only to meet another Australian.
Sorry, mate.
You're stuck with my Queensland accent.
Cherie is a commissioner for the Texas Forensic Science Commission,
which advises the state on misconduct in Texan crime labs.
And she told us that when you've got trace DNA evidence,
like from a knife blade or under fingernails,
you've got to extract DNA from the cells that they've left behind,
which means breaking the outside layer of the cell
called the lipid membrane.
Cells rupture pretty easily just with some basic chemistry and heating,
some detergents so that they can rupture the lipid membranes
and outspills all the DNA from the nucleus.
So once you've got the DNA on its own,
you can measure these very specific and special markers in it.
These are used to make a personalised DNA profile,
which kind of ends up looking like a bunch of little peaks
lined up on a mini
mountain range.
Then, this profile is compared to the profile of a known suspect, or compared to a database
of profiles.
And, if the profiles look the same, then it's a match.
Now, when you have a lot of cells with a lot of DNA, then DNA analysis is really reliable stuff.
This has been done for decades
on millions of samples in crime labs around the world.
A 2014 study of forensic DNA cases in the Netherlands
found that mistakes are made in less than 0.5% of the analyses.
Zzzz.
Zzzz.
Zzzz.
But Cherie tells us that if there are just a few cells in the sample,
like if you've just touched a knife, like in the Amanda Knox case,
or perhaps if the sample had been damaged
and you've just got those few cells to work with,
things can become a lot more complicated.
Because A, there's not enough DNA in the sample to start with,
or B, the DNA is there but it's severely degraded
so you'll get what's called a partial profile.
And a partial profile means that some of those peaks
that make up the mini mountain range of DNA are missing.
Even if you get a full profile from your trace DNA evidence,
you can still have problems.
Even if all the peaks are there,
problems can emerge because of a second issue.
When it comes to DNA analysis,
you have to copy the DNA
so that you can detect and analyse all the peaks.
Now, even if you have a lot of DNA to start with,
you still have to copy it.
But if you only start with just a small amount of DNA,
then you have to do more and more copying.
And when you copy it and you're starting with a crappy picture
and you're copying it and you're copying it
and you're copying it and you're copying it
and you're copying it,
the picture gets harder and harder to read.
Not only are you amplifying or increasing the signal
of the real amount of the DNA that's there,
but you're also increasing the noise.
You can end up with all of these unexpected things happening.
You can sometimes, you can run that same sample three times
and get slightly different results three times.
And so essentially it makes DNA interpretation a lot more difficult.
You can run a DNA profile and copy so much noise
that some peaks might just go missing
or peaks can pop up when they aren't
really in the DNA sample. So one way to sort this problem out, says Cherie, is that you can repeat
your DNA test several times and if you keep seeing the same peaks coming up then you know those peaks
are for real. So what does this look like when it plays out in a courtroom?
Well, this takes us back to Amanda Knox and Raphael Solicito.
To recap, the analyst who looked at this case originally
said that the knife, the alleged murder weapon,
which was found in Raphael's place, had DNA on it.
From Meredith Kirscher, the victim,
and Amanda Knox, the alleged murderer.
But two years later, the independent analysts who re-examined the evidence of this case
said that they couldn't conclude that Meredith's DNA was on the knife. So how did they get
different answers? Well, what was found on that knife were only tiny, tiny bits of DNA.
This was trace DNA.
And remember what Cherie said about needing to repeat
the matching experiment several times when the samples are small
so that you can keep seeing the same peaks coming up over and over again?
Well, the first analyst in the Amanda Knox case
was asked how many times she repeated her test.
And she said, only once.
But now, an independent analysis of that knife
by two court-appointed experts has found
that the DNA was so minuscule, so scant...
Conclusion.
Technology that allows scientists to extract DNA
from just a handful of cells
opens up a world of possibility for investigators.
But it also opens up a world of potential mistakes,
and investigators need to be ultra careful when interpreting this kind of evidence.
So that's trace DNA.
But another DNA issue that came up in the Amanda Knox trial
was about the DNA found on the bra clasp.
Kircher's bra clasp, found in the bedroom where she was murdered,
had Sollecito's DNA on it.
According to prosecutors, it was enough to convict them.
The Italian investigators matched some of the DNA
found on Meredith's bra clasp to Meredith,
which makes perfect sense.
It was her bra.
But there were more peaks in the DNA profile on the bra clasp
than just Meredith's.
And these extra peaks, the investigators said, matched Raphael.
This type of evidence where there's more than one person's DNA in a sample
is called a DNA mixture.
And it leads us to our next question.
Can you reliably untangle different people's DNA to figure out who's who?
So Cherie says that if there are just two people in the DNA mixture and you know one of their DNA profiles, say the victim,
then it's all pretty straightforward for scientists to eliminate the known DNA profile from the mix and leave behind that of the suspect.
But more than two people and things get messy pretty quickly.
Well, as soon as you get very, very complex mixtures into one sample,
that's when you're always going to have the most difficult time generating very clear results.
So, for example, when the Italian investigators
first analysed the profile of the DNA mixture on Meredith's bra clasp,
they identified peaks that matched Meredith's DNA
and Amanda Knox's boyfriend, Raphael.
But there were other peaks in that DNA mix as well.
The DNA analyst on the case, when asked about these peaks that didn't belong to Meredith
or Raphael, said in a preliminary hearing, quote, from my point of view, they're artefacts,
end quote.
In other words, she thought the extra peaks weren't from another person, but noise from
the copying machine.
But two years later, when the independent forensic experts
re-examined the DNA profile, they disagreed.
They thought those extra peaks weren't just noise,
but evidence that at least one other person's DNA was on that bra.
They called the initial assessment, quote,
not reliable, end quote. Another thing you need to know is that
DNA mixtures with tiny samples of DNA are very susceptible to contamination. When forensic
investigators turn up on a scene, if they're not careful, they can spread trace DNA through their
gloves when they touch doorknobs and knives or other evidence. That bra clasp wasn't found until 46 days after Meredith's body was discovered
and the independent experts wrote that contamination, quote,
cannot be ruled out, end quote.
The DNA evidence from the bra clasp and the knife were both thrown out,
leaving the prosecution with very little to go on.
Amanda and Raphael were acquitted.
So analysing DNA mixtures, it's challenging and things can go wrong,
which is why it was recently announced that more than 24,000 cases
where DNA mixtures were used in Texas would be re-examined
to see if there were any problems.
Basically, it's believed that Texan analysts
had been giving a higher degree of certainty that they were right in some cases, that they
could match someone's DNA within a mixture than they should have been.
Conclusion. DNA mixtures are hard to interpret when it's just left up to humans. In fact,
Cherie told us that in many cases, labs would not even try to interpret
a mixture with more than two or three people. Evidence from a crime scene with that mess of DNA
were once considered a dead end until... Enter complex statistical modelling.
A flurry of computer programs have hit the market claiming to untangle mixed DNA profiles with remarkable accuracy.
These programs weren't used in the Amanda Knox trial,
but since that case, more and more crime labs have started using them.
The FBI started using one of these programs at the end of last year.
All of these programs have technical-sounding names
like Armed Expert, True Allele, Lab Retriever, FST, and my personal favourite, Forensim.
One of the biggest programs in the business is Starmix.
It was partly developed in New Zealand.
Starmix assists with the rapid interpretation of complex DNA profiles. It does so by combining the most modern of sophisticated models
to interpret profiles in just minutes on a user's PC.
StarMix. Resolve more DNA mixtures.
What a sick beat, bro!
But there are very few independent studies
that have looked into how accurate and reliable
these computer programs are.
Our accents are completely different.
They're like, where's the car?
And we're like, where's the car?
As I was saying, there are very few independent studies that look into these programs.
Most studies that have tested their validity and reliability have actually been done by the companies who make them.
Still, Cherie says that many labs, including hers,
do their own tests to make sure that the computer program is spitting out reliable results.
And so she trusts them.
She says these programs are a helpful tool for her.
I can say that I would be much more comfortable going on the stand,
being able to give a statistical weighting
from one of
these products that we now use. Yeah, I'd feel more comfortable going on the stand with that
in my back pocket. A comprehensive report from the President's Council of Advisors on Science
and Technology concluded that even though we don't have many independent studies here,
from what we do know, these programs are better than human judgment and could produce reliable results, but only, quote,
under limited circumstances, end quote.
Like when only three people's DNA are in the mixture.
They said there is relatively little published evidence when it comes
to untangling cases where there are more than three people in the DNA mix.
More work needs to be done.
Here's Cherie. We'll still say that there are limitations than three people in the DNA mix. More work needs to be done. Here's Cherie.
We'll still say that there are limitations where these programs,
this should not be used to interpret so many person mixtures.
So, yeah, it doesn't solve all your problems.
It's not the magic bullet to fix everything.
Conclusion.
When scientists are analysing DNA,
the accuracy depends on how many people are in the sample.
With one person, it's easy.
If there are two people in the DNA mixture,
scientists can still usually tease them apart.
But if there's DNA from three or more people on, say, a bra clasp,
things can get really difficult.
And while computing programs are trying to help,
without independent testing, it's not clear how reliable they are.
When I asked Cherie about where this DNA evidence is all heading, she described it as a kind of arms race between what police want, what the public expects, and what scientists can provide.
So what happens is people find out, oh, we can get DNA from a doorknob.
So then all of a sudden the police now can get DNA from a doorknob.
So then all of a sudden the police now starts, you know,
swabbing every single doorknob that they can see.
And then, of course, now we've got lots of DNA, right?
And we've got all these different people, but they're small amounts.
So then what we do is say, okay, well, now we've got such small amounts of DNA, but we can't really see that well.
So now we've made our kits more sensitive.
Oh, yes, now we can pick up these really, really low amounts of DNA.
Oh, but now with increasing the sensitivity, now we've got more kits more sensitive. Oh, yes, now we can pick up these really, really low amounts of DNA. Oh, but now with increasing the sensitivity,
now we've got more background contamination coming through.
Now we have to...
And so then that creates another problem.
It goes round and round.
As we want to do more, we're able to do more,
but then that generates other problems that we then have to tackle.
Oh, technology.
It makes things better only to create new problems.
After the break,
we sniff out another forensic tool that's pushing the boundaries of science.
Welcome back to Science Versus. Okay, so we've solved the case of Amanda Knox.
You're welcome.
Now let's sleuth another incredibly high-profile court case,
the Casey Anthony trial.
This trial partly hinged on the ability to smell death.
For everyone outside of North America,
or if you were just listening to Firework,
on repeat for the entirety of 2011.
The Casey Anthony trial was one of the biggest courtroom dramas in America this decade.
Time magazine called it the social media trial of the century.
And it all focused around a young mother accused of killing her
two-year-old child. Some very troubling questions today in that Florida courtroom where Casey
Anthony sits on trial accused of killing her own little girl, Kaylee. Two-year-old Kaylee was
missing for a month before it was reported to the police. By the time Kaylee's body was found
near the family home, they'd been searching for her for five months.
And Casey, the mother, was the main suspect in the death of her daughter.
The case against Casey Anthony wasn't strong.
There was no slam-dunk physical evidence to pin the crime on Casey.
And so the prosecutors brought in some evidence at the edge of science,
including the alleged scent of a decomposing body in the trunk of Casey's car.
The smell of death, that is the gruesome topic this morning
as court gets back in session in the Casey Anthony murder trial.
The smell was first picked up by canine Gerris,
a German shepherd trained to detect the scent of a dead human.
The dog was taken so seriously that it was put on the stand.
No, the dog's handler, the sheriff's deputy, Jason Forgy,
was put on the stand.
He said that Garris was shown around Casey Anthony's car
and when the dog approached the trunk,
he did exactly what he was trained to do when he smelled human remains.
Garris comes out of the trunk with his front paws,
comes out of the trunk to the right rear passenger taillight bumper area
and gives me a final trained alert.
He goes into a down position.
The final trained alert to indicate that the trunk had a dead body inside.
Now, there was only one problem.
When they opened up Casey's car trunk, there was no dead body inside.
Kaylee's body was found five months later,
but the police said the dog's signal was evidence
that Kaylee's body had once been in the trunk,
alleging that Casey had put her dead body there.
So, what is going on here?
And can a dog really figure out where a dead body once was? Starting at the beginning,
when bodies decompose, they smell. And the fact that the bodies smell tells you that chemicals
are being released into the air that then lock onto our noses.
When we die, bacteria and fungi start infesting our bodies,
triggering a whole range of chemical processes that fart out molecules into the atmosphere,
and these are called volatile organic compounds.
So, for example, when bacteria eat carbohydrates in our body,
they can turn them into alcohols.
And as muscle and fat tissue degrades,
they can turn into other types of chemicals. Now, humans can't smell a lot of these compounds,
and particularly we can't smell them from very far away. But there's a superior species when
it comes to smelling. Chickens. No, dogs. Dogs. And they've been used to find dead bodies in clandestine graves for decades.
To find out how a dog like canine Garris in the Casey Anthony trial
was able to smell death in the boot of a car,
I met another dog trained to find dead bodies.
Zap.
It's a good panter.
Strong pant going on.
Zap is a black Labrador mutt,
and Matt Young, his handler,
says he's trained Zap to only search for the scent of human dead bodies.
So not dead pigs or dead chickens or a home-cooked meal.
Right, like, my game with Zap is he's looking for dead people,
he doesn't care if they're new dead people or old dead people
or girl dead people or boy dead people.
Dead people is dead people, and he goes and finds them.
We met Zap and Matt at one of their training sessions
in a scrubby wooded area in a park
about an hour's drive from the body farm in Huntsville.
The fireflies were out and the moon was starting to rise
and Matt had hidden some human placenta in the long grass around the park.
That great work?
Zap had to find it.
Zap, busca.
Yes, he said busca.
Zap speaks Spanish so that Matt doesn't accidentally give him commands when he's talking.
So Matt releases Zap and he runs straight for the patch of grass and starts barking.
Matt digs into the grass and voila,
the human placenta wrapped in plastic is revealed.
Good dog. Oh, good dog.
Now, Zap doesn't care if he's solving crimes or finding human remains.
He's really only after one thing.
A squeaky toy.
To train Zap to zoom in on human remains, Matt also hides animal decoys like dead pigs in the park. Now if Zap barks when he goes to the pigs but not the
human remains, he doesn't get his toy. Zap is so well trained that at one time he even smelled out
a body that had been dead and buried for 18 years.
But there's not a lot of research when it comes to how reliable these dogs are. A 2012 paper found
that the scientific literature on the capability of dogs trained to find human remains was, quote,
sparse, end quote. Still, it's pretty easy to tell if a dog can find a body
because the dog barks and voila, the body is there.
But using dogs the way canine Garris was used in the Casey Anthony trial
to find where a body maybe once was,
that's something harder to track.
For one, handlers can affect what their dogs do.
Researchers have found that when the handlers think they know where a target is,
it can influence their dog's behaviour.
It's something that Matt is keenly aware of.
He has to be very careful with his body language
because the slightest movement can give the game away.
Like, if Matt knows where the placenta is,
and then with his body language accidentally gives away
what he wants Zap to do.
Like, you're now just standing there leering at your dog,
and your dog's like, oh, yeah, this is totally it.
Which means the dog isn't necessarily smelling death,
it's just reading its handler.
Now, this issue came up in the Casey Anthony trial
because the defence argued that the dog handler
might have given canine Garris some indication
to do his gesture at the trunk of the car.
Now, since the police didn't videotape the search,
the jury had no way of knowing what body language he was using
and whether he did accidentally influence the behaviour of the dog.
And the thing is, we can't ask the dogs,
did you smell a dead body?
Why did you do what you just did?
Zap, doing an interview.
So when Zap indicates that he's found a dead body, he barks.
But he's a dog. He barks at other stuff too.
So what is he barking at now?
He's barking at that other dog. He's super jealous.
Can you tell the difference between a I'm jealous bark
and I found a dead body bark?
No, I can't.
I'm sure there is a difference,
but to me he's just a loud dog making a lot of noise.
Now, whilst there isn't a lot of research in this area of dogs generally,
there actually are a handful of studies
on whether dogs could sniff out where a body used to be.
And these studies do suggest that dogs might have this superpower. So in 2007, researchers in Hamburg placed two dead
bodies on a carpet for a few minutes and then they cut out the carpet and placed samples in glass
jars. They then mixed those jars with other jars that were full of clean carpet. They repeated the study in different ways and found that three highly trained dogs were
able to pick out the corpse-y smell in the contaminated carpet over 90% of the time.
And another study, published last year, found that around 86% of the time, dogs could pick out the smell of soil where a corpse had been up to 667
days after the body was removed.
Now that is impressive.
But still, it does mean the dogs were getting it wrong around 1 in ten times. And ultimately, if there's no body to confirm that the dog is right,
how do you know?
Here's Matt again.
Unless you actually are physically recovering anything from any case,
you don't actually know technically if your dog was right or wrong.
If somebody dumped a body in a location just on the ground for three days
and then came back, picked it up and moved it,
there's not going to be any evidence that we can readily find.
And so all you really are left with is law enforcement looking at you
and your dog barking at nothing going, well, these people are useless.
But of course, in the Casey Anthony trial,
the cops didn't say that Garris was useless.
He was used as a key piece of evidence.
Conclusion. Dogs are pretty remarkable at finding human bodies and could potentially find where a dead body once was. But when there
is no physical evidence in front of you, we can't be sure what the dog is trying to communicate.
Which is why prosecutors brought in a human person in the Casey Anthony trial
to back up what canine Garris, the dog person, had signalled.
An expert witness had never been allowed into a United States court
to say that they could detect the smell of a decomposing body
until, that is...
Dr Arpad Vass was called in as the smell of death expert to testify at the trial in 2011.
Here he is on the stand describing his initial reaction when he smelt the air from the back of the trunk.
Well, first I jumped back about two feet because the odour was pretty strong.
But it was, to me, the smell of human decomposition.
Arpad has been studying the chemicals
that make up the human smell of death for years.
He worked at America's first body farm at the University of Tennessee
and is one of the pioneers in this research field.
That's why he was contacted during the Casey Anthony investigation.
He was sent sections of the carpet lining of the trunk
and when he opened the
evidence cans. There was a horrible odor coming from that. It was very characteristic, in my
opinion, of human decomposition. He said it reminded him of something very specific. If you
have a potato that's, you know, in full rot, The potato is turning black and is squishy
and fluids are draining from it.
If you have a potato in that stage of decomposition,
I think the odor, at least to my nose,
is very similar to that of a human decompositional event.
And so how did you know that it wasn't just a rotten potato?
Well, one, there were no potatoes in the car,
and two, chemically looking, I mean, what my nose can sense
versus what the instrument tells me is there is completely different.
The instrument.
Arpad analysed the air emanating from the trunk liner
using established techniques like gas chromatography,
which pumps gas through a column
to separate out chemicals so then they can be identified.
So Arpad looked at the array of chemicals that were coming
out of the trunk liner and thought, it's a match.
Well, there was about an 80% match of what we considered
should have been there based on the age of the child or the decedent.
That is, Arpad said that he identified a scent that was 80% similar to that of a decomposing
body of a child. But the thing is, the trunk wasn't empty. It had a lot of junk in it.
No potatoes, but there was other stuff that really smelled.
And there was, you know, dirty diapers in there, there were empty pizza boxes, gas cans.
I mean, it was a huge plethora of material in the trunk of that car.
So Arpad had to analyse each of these things separately.
He had to get empty pizza boxes and diapers and other gas cans
and analyse all of those chemicals and separate them out
and then see whether they could have been responsible
for the smell in the trunk.
At one point, it was even suggested that the smell could have come
from a squirrel that the car had once hit.
So we actually looked at a squirrel as well.
We found a squirrel that had been hit on the road by a vehicle
and we allowed that to decompose on a similar piece of carpeting
as to what was in the trunk of the vehicle.
And he says that none of those things, the diaper, pizza, gas can,
emitted a chemical profile that looked like the smell of human death.
That profile was completely dissimilar to what we saw in human remains.
Now, central to Arpad's testimony and research
is that he knows the suite of chemicals that are uniquely emitted from a dead human body.
Well, there are about 500 different chemicals that emanate from a decompositional event.
We've narrowed it down to about 60 that are important.
And there are a number of different chemicals that are specific to human decompositional events, which we use as markers.
But here's the thing.
Most other scientists will tell you that we don't know for sure which chemicals are specific to human decomposition events
because a lot of chemicals found in human dead bodies
show up in other animals.
Part of the problem is that we don't know exactly
what mechanisms are farting out these particular odours, like exactly which bacteria is breaking down which cells in order to produce which chemicals.
We know the basics, that decomposing protein, carbohydrates and fat smell.
But pigs, squirrels, cows and humans, we're all made of that stuff.
So it's not too surprising that we emanate similar chemicals when we die.
And for this reason, researchers in this field are yet to find a consistent set of special chemicals
emitted from human corpses and not other animals. In 2012, Arpad wrote in a paper that he had found
four chemicals that, quote, appear to be human-specific, end quote.
But then, three years later, Belgian researchers found that three of those four chemicals that were human-specific
were found in other animals too.
Now, those researchers did say that they identified a bunch of chemicals
that were just in the decomposing humans in their experiment,
a unique cocktail.
But they pointed out that in order for that to be taken as fact,
more work would need to be done.
They wrote in their paper that this was a, quote,
first step, end quote.
That is, the results of the paper would need to be replicated.
So the research on the smell of death is still clearly unfolding.
Yet, four years ago, Arpad went on the stand in the Casey Anthony trial
and he said that he knew the markers that he had found
were definitely the smell of human death.
I asked him about that.
Wouldn't you feel more comfortable with your science
to know that someone else had replicated
it? Isn't that sort of the purpose of good science, that you have it replicated by different teams?
Who am I to say that one group needs to go and replicate my work? That's up to them.
I have no problem doing it. But more for the confidence in your own work,
so that you feel confident enough to be able to
go to court and say, this is the scientific consensus. Look at these teams that have
replicated the work. But if no one's willing to do that, what am I supposed to do about it?
Potentially not feel so confident about your work when you go into court.
No, no, that's not the right answer.
What is the right answer?
The right answer is to be confident in your work
and if someone wants to replicate it, go for it.
Go ahead, replicate it.
Other scientists criticised Arpad's work during the Casey Anthony trial.
They said he came out of the gate too early with this one.
Dr Donovan Haynes is a chemist at Sam Houston State University,
and he also studies the chemicals that get emitted when animals die.
Donovan remembers the scientific community's reaction
to Arpad Vass's research being used on the stand.
In that Vass case, if you look,
there's a lot of discussion among forensic scientists
about should it have been allowed in court at all?
And that's them erring on the side of saying the bar is pretty high here before we put something in front of a jury that's going to potentially convict somebody to make sure that that science really is good vetted science in that case. Why do you think that some in the forensic science community
and some scientists were not happy with the testimony
and with the research you presented in court?
That's an interesting question.
I really don't know the answer to that.
Some of it possibly is jealousy.
Some of it could be perhaps they've been working on
and couldn't come up with a suite of chemicals.
I honestly don't know the answer to that.
The jury in the Casey Anthony trial deliberated for 10 hours and 40 minutes
before giving a verdict.
State of Florida versus Casey Marie Anthony.
As to the charge of first-degree murder,
we, the jury, find the defendant not guilty,
so say we all. Dated at Orlando, Orange County, Florida.
Conclusion. Using chemical analysis to detect where a dead body has been
is in its very early stages of research. Currently, the studies are inconsistent
and there is not a reliable suite of chemicals that can be measured from decomposing human bodies.
So when it comes to science versus emerging forensic sciences,
do they stack up?
When you have a single sample with a lot of DNA in it,
the evidence here is solid.
DNA analysis works and can reliably pin a perp to a crime.
But when there is only a little bit of
DNA evidence in your sample, only a few cells, the results can get less reliable. As you try to copy
the DNA, noise gets in the analysis and it's harder to interpret. Another problem, when there's
several people's DNA in one mixture, things get really complicated.
One new solution is to use computer programs to untangle the mess. Now, these programs show a lot
of promise, but right now we don't have much independent research into their reliability.
As for the smell of death, dogs have been used for a long time to lead search parties to corpses in unmarked graves.
Trouble hits when the dogs are used to find where bodies used to be.
Because dogs don't talk, we can never 100% know
what they're really trying to say.
What's that?
Nah, I know you might be able to do it, it's just that...
Oh, just take the toy. at? Nah, I know you might be able to do it, it's just that...
Oh, just take the toy.
As for people knowing the smell of death, the science here is inconsistent and we need more work.
So, the big question now. Why is
experimental science entering the courtroom and being presented
as scientific fact.
Here's what Arpad had to say about it.
I mean, if the court's willing to accept my expertise and my testimony,
that's up to the court to decide.
It's not up to me to decide.
I do the research, I present the results.
If the court wants to say it's acceptable for the legal system,
then that's the court's problem, not mine.
Arpad says this because there are hearings before a trial begins
where judges decide if science is valid enough to be presented in the courts.
The problem is, of course, that judges and lawyers
often don't have any scientific training.
Deciding what is a valid science often involves lawyers just throwing their experts at judges.
This is hardly a peer-reviewed process.
So if we're not going to throw out this system, then we'll need forensic scientists, as well as cops and lawyers, to step up. When it comes to the science, Dr. Joan, by the way, from the body farm,
she says she knows there have been problems with her field.
But she said she's also worked with a lot of good forensic scientists
who are just working hard to solve cases.
And ultimately, she's hopeful about where this field is headed.
My perspective over the years of teaching this for 20 years and working with
law enforcement for 20 years in this, have there ever been mistakes in a courtroom? Absolutely.
Absolutely. It's not by the fault of deception. It's just by the fault of we're growing,
we're learning, we're every day, we're trying to become better at what we're doing. And that is the goal of all forensic scientists.
And if it isn't, they are ferreted out.
Eventually they are ferreted out.
And there are calls for more ferreting to be done.
In an editorial published in an academic journal last year,
which read more like a call to arms,
forensic scientists at the University of Dundee wrote,
quote, we must be both bold and brave about exposing and recognising baseless junk science
and cowboy practitioners, end quote. And exposing those cowboys is important because when science
is admitted into the courts before it's rigorously tested. People might be wrongly convicted, victims can be misled
and guilty people can walk free.
That's science versus forensic science.
This episode has been produced by Diane Wu,
Shruti Ravindran and Heather Rogers.
Our senior producer is Caitlin Sori.
Edited by Annie-Rose Strasser and Caitlin Kenny.
Fact-checking by Michelle Harris.
And a big thanks to Joseph Lavelle-Wilson,
Will Doolin and Beth McMullen.
Thank you for helping us out.
Sound design and music production by Matthew Boll,
mixed by Martin Peralta and Matthew Boll.
Music written by Bobby Lord and Martin Peralta.
Next week is our final episode for this season
and we're tackling antidepressants.
Do they work any better than a dummy pill?
Frog sperm and feathers and fur and spiders
that were given as treatments that people would ingest
and if it didn't kill them, it might make them better.
And that's part of the history of the placebo effect.
I'm Wendy Zuckerman.
Back to you next time.