The Smell of the Blue Ball

A favorite activity of the Case dogs is the “Find It” game. We play this out in the training building and begin by asking the dogs “What’s Hiding Today?”. We all visit the giant toy bin and select a toy for the day’s game. I show the toy to everyone, making sure that each dog gets a good long sniff. The dogs then run to the storage room where they wait as I hide the toy. The door is opened, dogs burst forth into the room and the competition begins! Everyone races around, air scenting, searching every nook and cranny of the room. The first dog to find the toy is the winner, grabbing the toy and running with it to home base to get their prize, a yummy treat.

Mike and I have played this game with our dogs for more than 30 years. While we have seen a range of talent and search strategies among them, all of our dogs adore this game, turn cartwheels to play it and always ask for “one more round” no matter how often we play. In addition to being convinced of the pleasure that dogs take in searching, it has always been obvious that our dogs not only search for the toy using primarily scent (olfaction), but also that they are able to discriminate between toys and will select only that toy that is chosen for the day’s game. For example, if another toy, one that was not chosen, has been left out somewhere in the building, it is summarily ignored, even if that toy has been hidden for the game on a previous day. It has always seemed that our dogs not only have been searching using their noses (not a big surprise of course), but that they are keenly aware of the smell of the blue ball versus that of the red tug toy.


Picture a Blue Ball: The mental representations of objects is something that most of us take for granted. For example, think of your dog’s favorite toy; for my dogs, this is a blue Planet Dog ball. What mental representation do you conjure? I bet it is a visual representation, correct? (I see the blue ball in my inner brain right now). We may naturally assume that other animals represent their worlds similarly. However, for species such as dogs, whose strongest sense is olfaction, it is quite possible, expected really, that their mental representations of objects may be more strongly olfactory than visual – in other words, the smell of the blue ball.

A recent study examines just this question – Do dogs represent objects as odors?

The Study: Researchers at the Max Planck Institute in Germany used a study paradigm called the “violation of expectation” test. There are various ways to design this test, and in this particular set-up, dogs were asked to track the odor trail of a selected toy. At the end of the odor trail, they found either the expected toy or, big surprise, a different toy! The researchers predicted that if the dogs had an internal olfactory representation of  the selected toy, then finding a different toy at the end of its scent trail would be unexpected and discounted – meaning that they would hesitate or continue to search for the selected toy – just like my dogs do in the find it game. Conversely, if the dog simply grabbed the unexpected toy, this would suggest that although dogs are highly capable of following a scent trail (something that is already well supported and known), they do not have a strong internal scent representation of the scent of a specific object. Given what we know about the dog’s extraordinary sense of smell, the researchers hypothesized that dogs do indeed have very specific and distinct odor representations of objects and that the first scenario would predominate in their test.

The Dogs: They tested a group of 48 adult dogs of varying breeds and with different training backgrounds. Half of the dogs were trained either in search and rescue or police work. The other half were dogs living as family pets who had received no formal training. For each dog, two “high interest” toys of similar size were selected for use in the search tests. Each dog was tested in four randomized conditions in which one toy was dragged along the floor to leave a scent trail and then hidden in a cupboard. The routes and the hiding places were varied and dogs found either the expected toy (i.e. the toy that left the scent trail) or the unexpected toy, in two trials each. Each  dog’s searching and sniffing behaviors, time to find the hidden toy, and response when finding the expected or unexpected toy were recorded.

Results: All of the dogs searched for and found the toy successfully within 2 minutes and the majority retrieved the toy after finding it. Sniffing behavior was used to search in the majority of searches (75 %), and most dogs used both air scenting and ground sniffing to find the hidden toy. Dogs found the hidden toy significantly faster when they sniffed versus the smaller number of trials in which they attempted to find the toy visually. Here are other interesting results:

  • Surprise! On the first trial, there was no difference in the amount of time that it took to find the expected versus the unexpected toy, but significantly more dogs hesitated to retrieve the unexpected toy compared to the expected toy. This suggests that the dogs experienced a “violation of expectation”, supporting the hypothesis that they had an internal scent representation of the specific toy.
  • Working dogs vs. pet dogs: On the very first trial, the working dogs searched at higher speeds and were considerably faster at finding the hidden toy than the pet dogs. However, after the first trial (when it is assumed that the dogs “learned the game”), there was no difference in search speed or success between the trained working dogs and the pet dogs.
  • Not just the sniff: Dogs used a combination of both sniffing and visual searching in most trials. Interestingly, they tended to use vision immediately and when the toy was not obvious, such as when it was hidden in peripheral spots, they relied more heavily upon sniffing.

Take Away for Dog Folks: This study provides confirming evidence for something that many of us witness daily with our dogs – that they identify and discriminate among different objects using their sense of smell. But, there is also a bit more here. This study targets cognitive questions about internal representation and how human perceptions, which are primarily visual, may differ dramatically from a dog’s perceptions and representations. When Ally and Cooper compete to find the hidden blue ball, they may know what that blue ball looks like, but the internal image that they have of it is probably not an image at all…..but rather is a smell. The smell of the blue ball.

Can we, as a primate species, even know  what that is like? Probably not, but it sure is cool to attempt to understand it.

Happy Training!

Cited Study: Brauer J and Belger J. A ball is not a kong: Representation and search behavior in domestic dogs (Canis familiaris) of different education. Journal of Comparative Psychology 2018;

Interested in Learning More about Canine Nutrition? Take a Look at the New Science Dog Course!

“Basics of Canine Nutrition”


Interested in learning more about canine cognition and training? See Linda Case’s 2019 book – “Dog Smart: Evidence-based Training with The Science Dog“!

8 thoughts on “The Smell of the Blue Ball

  1. Alternatively, there were times when he would poop in a pile of leaves, and in some cases, after I got my baggie out, I couldn’t make out where his poop was located. I would say, “Where did you go?” And he’d look at me, then go over and sniff his scat and I’d pick it up.

    This is inline with: “Dogs used a combination of both sniffing and visual searching in most trials. Interestingly, they tended to use vision immediately and when the toy was not obvious, such as when it was hidden in peripheral spots, they relied more heavily upon sniffing.”

    Liked by 1 person

    • Hi Lee, I completely agree, and though I cannot speak for the researchers, I doubt that they would maintain that dogs lack visual imagery (just as we do not lack olfactory representations – I can definitely conjure up the smell of a good cup of coffee in the morning! 🙂 ). Rather, my impression from this research is that it attempts to tease out how important and specific olfactory representations may be for dogs and to determine how strongly such representations (if they exist) influence a dog’s behavior and decision making. Like your Dalmatian, my guys definitely use vision to search and my dogs definitely check in visually with us when they are off during walks in the park. What I found interesting in this study is that it is the first, that I have ready anyway, that tries to specifically address the question of how a dog represents specific objects mentally. While this is just a bit of data to address that question (see Peter’s note above), I think it was a clever study design and interesting – including the bit about pet dogs stepping up very quickly and performing as well as the trained dogs! Thanks for posting! Linda


  2. I’m not sure your conclusion is correct.

    “When Ally and Cooper compete to find the hidden blue ball, they may know what that blue ball looks like, but the internal image that they have of it is probably not an image at all…..but rather is a smell. The smell of the blue ball.”

    I think dogs *do* have mental images. Many years ago while walking my Dalmatian in Central Park I would let him walk ahead of me on the path we were taking. I had trained him to always look back to see where I was every ten seconds or so, which he did quite reliably.

    Then at some point I tried “tricking” him by veering a little ways off the path. Whenever I did that and he looked back at the point where he expected me to be, he would look surprised, then scan the area to find me again.

    This suggested to me that he had a mental picture (internal image) in his mind of where he expected to me to be. There was no olfactory component at all.

    Liked by 1 person

  3. I wonder what expectation was being violated ? Dogs surely expect scent trails to stay about the same along their lengths – here they experienced a sudden change in scent right at the end. Separating an olfactory representation of the object in their heads from an expectation that scent trails stay the same would need some fancier experiments than this. A first step would be see how dogs respond when a drag trail is started with one object and continued with another at some point along its length, with a sharp transition, or with an overlap.

    Liked by 1 person

    • Hi Peter, I don’t think the researchers were claiming that this particular study design proves that dogs possess an olfactory representation of objects, but rather that it provides some evidence to support this claim. The design that they used is related to “violation of expectation” tests used with children and primate species (I think first developed by Piaget, but not sure), that test for an understand of object permanence. In those tests, an object disappears behind an opaque screen as the child (or non-human animal) tracks the object visually. Some of the time, the same object appears on the opposite end of the screen, and at other times, a new object appears. If the child shows surprise (measured by increased gaze, I believe), that is taken as support for an understanding of object permanence. It seemed that the researchers in this paper were designing a (imho, rather clever! 🙂 ), approach to testing for this concept as an internal olfactory representation in dogs (similar to what Alexandra Horowitz has done with her dog pee studies). I love your idea for an additional approach to this question and think that several study designs certainly could be used to try to tease out exactly what types of concepts dogs possess mentally. Thanks (as always) for writing in! Linda


  4. Five years ago I bought a dog for our police department so they let me tag along on training days. The dogs would track a person and be rewarded with a bite-they would engage with the other handler who was wearing a protective bite sleeve. After a policy change, they started leaving the dog’s toy at the end of the track. I don’t know if the dogs were “surprised” by the change, but the toys are already high value objects and the dogs continue to work admirably.

    My personal dog, a 4.5 yo “pet” Doberman is not a power chewer like the police dogs. His favorite is a glow ball–tennis ball size, hollow interior and a hole at either end so he collapses it while chewing. They eventually split so I keep spares. For awhile he had two at once and definitely preferred the older one with the small split. When both were under the furniture, he would consistently try to get his fave, sometimes even refusing the “wrong” one if I got that one out for him. He certainly wasn’t working off of visual cues.

    I love watching dogs do nose work. It looks like magic!

    Liked by 1 person

    • Hi Lelia, Your story about the police dogs is interesting! I too wonder what the dogs thought when they first found their favorite toy rather than the person who they were tracking – especially if they had been trained with the person as the reward! Love your description of your own dog’s torn glow ball. One of our instructors at AutumnGold has a Vizsla, Gracie, who is similarly particular and knows her special ball from all other toys. Totally agree about nosework – I never tire of watching dogs find things via scent!


Have a comment? Feel welcome to participate!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s