What Makes a Detection Dog?
There are three things that make a detection dog; hunt drive, odour recognition and the response.
Hunt DriveIn every dog exists motivation to hunt and search. Whether it be for the pursuit of food, prey or for the opportunity to breed. However, that doesn’t mean that we can tap into that drive when we want, where we want and (most importantly) for what we want.
There are typically two things that factor into hunt drive;
1. How rewarding the search itself is, and;
2. How much the dog wants the reward they get for finding the magic smell.
Dogs that have been selectively bred for detection work have both in excess; they’d happily search until they drop dead, and they just about have a heart attack when they get their reward.
Dogs that haven’t been selectively bred for the task have varying degrees of both, but we can directly influence them both through careful training and the selection of a high value reward.
The reward is everything; if we have something they want, we can condition them to think that the act of searching for that reward is the best thing ever.
A dog with odour recognition has been trained to recognize the target odour as being very valuable. In scientific terms, it has been classically conditioned in that it has gone from something that the dog has no value for (positive or negative), to something that gets them very excited when they smell it. In layman's terms, the target odour is a pay check; when he finds it, he cashes it in for his reward.
There are two parts to odour recognition; knowing what odour pays and what odour doesn’t. The target odour can be pretty much anything, from essential oils to animal scat. It’s essential that whatever odour is being used, it is isolated as the only relevant odour in the environment. So, if the odour is held by a cotton swab, or it’s housed in a cardboard box, PVC pipe, metal tin etc. there should be numerous other identical, sterile items in the search area too. This not only isolates the target odour as being solely relevant, but it also reduces the chances that the dog will start relying on any visual indicators as to the location of the scent.
As a part of training, the dog should be exposed to and proofed on non-target odours; distractions and similar scents. The usual distractions that are proofed against are food and toys. Odours that are similar to the target scent isn’t something that Nosework participants really need to worry about. Where it’s most relevant is in venues like conservation detection. For example, I’ve worked with a dog that was being trained to find Porcupine Scat. As part of his training, we brought in the scat of other animals that ate a similar diet and lived in the same region, to ensure he knew exactly what he was searching for.
Proofing on these non-target odours means that we’ve ensured that they will not offer their trained response, but it does not mean that they won’t investigate them. It’s normal for them to pay attention to something interesting and novel in the environment. What matters is that they don’t alert to it or try to access it.
The dog’s response is its way of communicating to its handler that it has sourced the target odour. Also called an alert or indication, the exact behaviour that the dog offers differs a great deal according to its application, the training method used and handler/trainer preferences. All response behaviours can be lumped into two broad categories;
An active alert is one where the dog directly interacts with the training aid or the source of the odour. They may bite it or scratch it, or both. Active alerts are the easiest train, because it generally involves capturing a dog’s natural response when trying to reach something they value.
There are a number of reasons why active alerts are highly undesired, regardless of the application;
1. It can damage property. The vehicle search in some Nosework sporting organizations is the main place where this will rear its ugly head. The potential for damage that scratching causes is enough for people to decide against volunteering their cars for searches during competition and training.
2. It contaminates the search area. If a dog scratches, pushes or bites a box in the container search, he’s impacted the search of every dog that runs after it. Many trial organizers keep spare hot boxes to replace those that have been damaged to an almost 1-1 ratio. But the tampering can potentially spread odour all over the area too.
3. Because it is a natural response, some dogs may offer it if they’re interested in something else in the environment. The most common place this occurs is when intentional distractions are used in competition. The dogs are so used to saying ‘hey, there’s something in there that I want!’ that they at least try to get away with it.
4. It can be hazardous to health. Ingestion of essential oils is generally an undesirable habit, but it can also be harmful. For narcotics dogs, the ingestion of any quantity of drugs is life threatening. Yet it’s so common amongst working K9s with an active alert that many handlers keep emergency drug odour kits on hand. It goes without saying that a bomb dog scratching a ‘hide’ is not only hazardous to the dog, but to everyone around it.
*Barking is often considered an ‘active’ alert, even though there’s no direct interaction with odour. A bark alert is typical for a search and rescue dog, and the above problems do not apply.
A passive response is one in which the dog does not interact with the odour source or training aid. The most common passive alerts are the focused response, the sit or down, or a combination of both.
The Focused Response
This is the most accurate response. The dog works to the highest concentration of odour and freezes with his nose at that point; like a bird dog pointing game.
For Nosework competition, this accuracy is very beneficial. It’s within the judge’s power to ask handlers where the hide is located, while their dog is responding. To which the answer simply is; ‘where the dog’s nose is!’ Having the criteria be so narrow and specific also carries over to more challenging hides that are out of reach. Focused response dogs are generally very accurate, even in this scenario. Dogs that have been paid for responding inches away from source when it’s reachable, will tend to be considerably less accurate when source is unreachable.
For all its strengths, the focused response is unnecessary for most detection applications. A less precise response to drugs or explosives gives the handler enough information to act accordingly. For the narcotics dog, it gives probable cause to search an item or location. For the bomb dog, it means it’s time to get the hell out of the area. There's no ‘extra points’ for having a dog alert to the exact location, to within a millimetre; a zipper on a suitcase instead of the suitcase itself.
The biggest drawback to the focused response is that it can be very challenging to teach and maintain. There have been numerous methods popping up over the last few years aiming to refine a broadly applicable system of training it, with varying results.
Probably the most common response in working detection dogs is the sit-stay-pay, where the dog sources odour, sits and looks to their handler for a reward. It is very reliable, obvious and compared to the focused response, simple to teach. Most working detection roles which will be encountering odour low down – such as with conservation or truffle detection dogs – the sit is replaced with a down.
The main problem with this response is that it isn’t accurate. Because dogs are being paid when they’re looking at the handler, many tend to get sloppy about sourcing the highest concentration of odour. They know the shortest route to their reward is sitting, so they do it when they’re in the vicinity of source. Some handlers combat this by asking; ‘show me’, a cue which induces the dog to re-source the hide and alert again.
Accuracy isn’t the end of the world for real world detection. The sit/down response is detection’s equivalent of a functional, everyday heel. While the focused response is like competition focused heeling. The latter is prettier, and highlights the trainer’s skill better, but it is not needed to get the job done. (Damn, for someone that exclusively trains the focused response, I’m really not much of a salesperson!)
The combination of a focused response with a sit/down is another acceptable alert. Sometimes called the deferred-final-response or sit-and-stare, this was actually the progenitor of the focused response. As far as I know, it was first developed for gas-leak detection dogs, then mine detection dogs at Lackland Air Force Base, until it caught on in the explosives detection department. Sit-and-stare dogs will source odour, then rock back into a sit (or fold in to a down) and stare at source. It is more accurate than sit-stay-pay response, but due to the distance that assuming the sit creates, it’s not as accurate as the focused response.
Training this behaviour requires an additional step after training the focused response, so it can be argued that it’s the most challenging of all responses to teach.