"So many places to hide a dead body."

That’s what my mom remembers thinking on her first drive cross country during honeymoon number one. Maybe this was a premonition of things to come — marriage number one was short-lived — or maybe this was the only observation a person holed up in Queens, New York could make when seeing the cornfields of Iowa for the first time.

While I get where my mom's coming from, I think she’s missing a critical point. Sure, there are lots of places in America to hide a body, but maybe not without a trace. In addition to the many criminal-catching techniques — DNA, video cameras, cell phone triangulation, and super-policing robots — there are also dogs.

Cadaver detection dogs, human remains detection dogs, victim recovery dogs — no matter how you slice it, they all do the same work: find human remains and alert their human handler. It’s important, challenging work that Cat Warren tackles in “What the Dog Knows: Scent, Science, and the Amazing Ways Dogs Perceive the World,” now out in paperback, and covered in last week’s Dog Spies post.

But can cadaver dogs work without a body? For example, what happens if a body had been somewhere but is no longer there? Would a trained cadaver detection dog be able to say, "Hey! I know a body’s not here right now, but one sure was here!" This was the question that Lars Oesterhelweg and colleagues posed in a 2008 paper, and this is the case that prompted their question:

“A married couple went on a sailing trip. During this trip, the wife disappeared and was reported missing by her husband. A criminal investigation was initiated and the husband was under suspicion of having murdered his wife. A cadaver dog of the State Police of Hamburg was ordered to search the yacht and gave a signal on a mattress in the bedroom of the yacht.”

So what’s the answer? Could a cadaver detection dog indicate that a dead body had been there? And if so, how long would the body need to be on the surface for the dog to pick up the scent? While the postmortem interval (how long a person had been dead before making contact with the surface) could play into the equation, here’s how Oesterhelweg and colleagues went about their inquiry: two recently deceased individuals were brought to the University Medical Center in Hamburg where they were wrapped in a new, cotton blanket and then placed on small carpet squares for different amounts of time — 10 minutes and 2 minutes. These carpet squares would serve as “the medium for the odor transport." Importantly, control samples were also collected; these came from live people who lay on different carpet pieces. How did someone get the role of ‘live person’ in this study? They had to “[deny] having had any contact with deceased tissues…” Good to know.

Enter the three reasons not to leave a dead body on a carpet. Three trained cadaver dogs were brought in to sniff the carpet pieces (see Figure 2). Would the dogs alert (scratch or bark) when they smelled 'deceased person' on a carpet, and not alert to the smell of 'live person'?

Successful noses?

The cadaver dogs did their job and overwhelmingly indicated on the carpet pieces that smelled of a deceased person. They performed slightly better when the body had rested on the carpet for more time, 10 minutes as opposed to 2 minutes.

But the dogs were not always right. Over the course of 354 searches, they occasionally gave false signals, either "over-runs" — not alerting on a carpet impregnated with the dead body — or "mis-signals" — alerting on a carpet devoid of a dead body. Even so, false signals were relatively uncommon. In fact, most errors came from one of the three dogs, a point Cat Warren picks up on in “What the Dog Knows”“That’s the reality of working dogs. A few are excellent, while some are very good. (Others are horrid. The last category wasn’t represented in the study).”

There you have it — three trained cadaver dogs did a very good job indicating, “Yup! I know a body’s not here right now, but one sure was here!" While not totally infallible (of course, neither are humans), the dogs gave very few false alerts. Go dogs.

But is “very good” good enough when we are talking about a possible homicide case? (I should hope not). The researchers offer that “a positive signal by a trained cadaver dog should not be used as the sole evidentiary piece in court,” especially when there’s no body. Instead, dogs are an excellent tool which law enforcement can follow up on. For example, “The reliability and accuracy of the dog’s signal may be further enhanced if at least two or more trained cadaver dog/handler teams are used independently at one crime scene.”

Two dogs? Anyone else see an opportunity for a new dog-cop buddy movie?

~ ~ ~

This is the third post in the March 2015 Dog Spies Nose-a-thon. Also in this series, Part 1: This Month, Step Inside the Dog’s Nose, and Part 2: Dog of the Dead: The Science of Canine Cadaver Detection. And related, One Day You Will Smell Like A Dead Chicken, and Would Your Dog Make a Good Cadaver Detection Dog?

Images: Brandon Anderson Crime Scene Tape, Flickr Creative Commons License; Figure 2. Oesterhelweg et al. (2008). Forensic Science International.


Oesterhelweg L., K. Rottmann, J. Willhöft, C. Braun, N. Thies, K. Püschel, J. Silkenath & A. Gehl (2008). Cadaver dogs—A study on detection of contaminated carpet squares, Forensic Science International, 174 (1) 35-39. DOI: http://dx.doi.org/10.1016/j.forsciint.2007.02.031