Can Dogs Smell Quantity of Food?

Can dogs tell how much food there is just by using their nose?

By Zazie Todd, PhD

We all know dogs’ noses are amazing. From careful attention to the pee-points on their walk, to working as drug or explosive detection dogs, it’s clear dogs have an excellent sense of smell. So it’s surprising that most studies of olfaction are about specially trained dogs, and less attention has been paid to the average pet dog. A paper in press in the journal Learning and Motivation, by Alexandra Horowitz, Julie Hecht and Alexandra Dedrick, sets out to change all that by asking, can dogs smell whether a closed plate contains a small or large quantity of food?

The research is based on a study of canine preferences by Prato-Previde et al in 2008. They found that when dogs could see two plates, one with a small amount of food and the other with a large amount of food, they preferred the plate with more food. No surprises there! However, when given the same choice but with their owners making a fuss over the plate with a small amount of food, most dogs changed their preference and went for the plate their owner was interested in.

The study by Horowitz et al replicates this paper, but using closed plates so that dogs could not see the quantity of food; instead they had to rely on their sense of smell. One plate had only one piece of hot dog whereas the other had five pieces. The plates were prepared beforehand and ‘closed’ by putting another plate on top, so the contents were not visible. They were then given to the experimenter so that she did not know which plate had the most food. Pieces of dried liver were used for one participant who did not like hot dog.

Sixty-nine pet dogs took part, although five were excluded from the results (e.g. because they always chose the plate on one particular side). To ensure they would be interested in food, their owners did not feed them during the three hours prior to the experiment. 

The study took place in a small empty room, either at the University or at a doggy daycare. To get the dog ‘warmed up’, they were shown a plate on the floor with a piece of hot dog, and allowed to take it. There were three experimental conditions, and dogs took part in each one in turn (only fifty-three dogs did condition two). The sessions were video-recorded for later analysis.

In the first condition, the owner sat on a chair with their eyes shut. They held the dog by its collar. The experimenter approached with two closed plates, got the dog’s attention, and allowed it to sniff each plate for about three seconds. Then she put the plates on the floor and turned her back.  The owner released the dog. When it approached one of the plates, the experimenter turned around and threw a piece of hot dog as a reward. 

The second condition was very similar, except the owner’s place was taken by one of the scientists. The plates were prepared as before, and presented by the experimenter. Then, once she had placed both plates on the floor, the owner picked up the plate with the smaller item of food and said to her dog something like “Oh boy, this is good food, yummy!”.  Then they put the plate back, went and stood next to their dog, and the dog was released. As in the first condition, when the dog approached a plate it was given a piece of hot dog.

Although 61% of the dogs approached the plate with the large amount of food first, it wasn’t significantly different from chance.

However, when the owner talked happily about the plate with a small amount of food, the dogs’ preference changed. Dogs were significantly more likely to choose the smaller amount when their owner was enthusiastic about this plate.

Forty-four of the dogs paid more attention to one plate than the other; amongst these dogs, it was the plate containing more food that received significantly more attention. This suggests they could detect something, even though it didn't influence their choices.

The third and final condition investigated the effects of adding a scent to the plate with the most food. The scents were from products aimed at the pet market, and hence safe for dogs: mint from a spray designed to improve the smell of a dog’s breath, lavender from a doggy deodorizing spray, and distilled vinegar since this might be used in cleaning. This time the plates were open, so the dog could see the quantities of food, but otherwise it was the same as the first condition. 

Did the addition of a strong scent to the plate with the large amount of food change a dog’s preference? You bet! 64% of the dogs now preferred the small, unadulterated, amount of food for at least one of the smells.  There were individual differences between the dogs as to which of the smells they did not seem to like.

The design of this experiment is really nice because great care was taken to ensure there were no biases that might affect the results. The authors also checked that there was no difference between simultaneous versus alternate presentation of the two plates.

Whether dogs went to the smaller- or larger-food plate first, they were rewarded with one piece of hot dog. It would be really interesting to see a future study in which the size of the reward matched the choice of the plate, to see if the dog would learn over time which choice earned the larger reward.

So, contrary to an earlier study based on sight, dogs could not detect the larger amount of food by smell alone. However, dogs were influenced by their owners when they made a fuss of the smaller-food plate. Adding a strong scent to one of the plates also caused dogs to change their preference. This study is a fascinating step in our understanding of the olfactory abilities of ordinary pet dogs.

What is your dog’s sense of smell like?

References:

Horowitz, A., Hecht, J., & Dedrick, A. (2013). Smelling more or less: Investigating the olfactory experience of the domestic dog. Learning and motivation, 44(4), 207-217.

Prato-Previde, E., Marshall-Pescini, S., & Valsecchi, P. (2008). Is your choice my choice? The owners’ effect on pet dogs’(Canis lupus familiaris) performance in a food choice task. Animal Cognition, 11, 167-174.