Researchers prove dogs are able to differentiate colors

July 25, 2013 by Bob Yirka report
A yellow labrador retriver dog with pink nose. Credit: Wikipedia.

A team of researchers in Russia has conducted a series of experiments that prove that dogs are able to distinguish between different colors. In their paper published in Proceedings of the Royal Society B, the team describes the experiments with dogs they conducted and the results they found.

For much of history, have been assumed to be able to see only in black and white—their ability to differentiate between different colored objects was believed to be due to differences in brightness. In this new research, the team in Russia built on research recently conducted in the U.S. that found that dogs have two cones in their eyes suggesting they should have some ability to differentiate . Humans as most remember from grade school, have three cones, which allows for seeing all three primary colors. Since dogs have only two, they should be able to see some colors, but not others—blues, greens and yellows, for example, but not reds or oranges.

To find out if dogs are in fact able to see colors and to distinguish between them, the team conducted a clever experiment. First they trained several dogs to respond to one of four different colored pieces of paper: light or dark yellow and light or dark blue (by putting paper pairs in front of feedboxes that contained meat.) The dogs soon learned that certain colors meant they were in for a treat.

Next, using the same dogs that had been trained to respond to certain colors, the team placed pieces of paper with the color that they'd been taught to respond to in front of a feed box, along with another piece of that was brighter, but of a different color—a dog trained to respond to light blue for example would hopefully respond to dark blue instead of light yellow. The researchers found that a majority of the dogs went for the color identifier rather than brightness identifier most of the time, proving that they were able to distinguish color and were not relying on brightness to find their food treat.

The researchers suggest their findings indicate that most animals with just two cones are likely able to differentiate between colors and thus it's likely they respond in ways that have not been previously studied.

Explore further: Dogs may understand human point of view, researcher finds

More information: Proceedings of the Royal Society B doi: 10.1098/rspb.2013.1356

Related Stories

Seeing color traced back to genetic mutations

May 21, 2012

From the inside of our heads, it feels as if colors are intrinsic aspects of the outside world and our eyes are beautifully designed to see them. But we humans are merely sampling the possible ways of sensing the spectrum ...

See spot see

March 2, 2013

(HealthDay)—It's a dog-see-dog world. With no sniffing involved, dogs can recognize the faces of other dogs among the faces of humans and other animal species, according to a new study.

Recommended for you

Neanderthal DNA contributes to human gene expression

February 23, 2017

The last Neanderthal died 40,000 years ago, but much of their genome lives on, in bits and pieces, through modern humans. The impact of Neanderthals' genetic contribution has been uncertain: Do these snippets affect our genome's ...

New gene for atrazine resistance identified in waterhemp

February 23, 2017

Waterhemp has been locked in an arms race with farmers for decades. Nearly every time farmers attack the weed with a new herbicide, waterhemp becomes resistant to it, reducing or eliminating the efficacy of the chemical. ...

Arctic 'doomsday' seed vault receives 50,000 new deposits

February 23, 2017

Nearly 10 years after a "doomsday" seed vault opened on an Arctic island, some 50,000 new samples from seed collections around the world have been deposited in the world's largest repository built to safeguard against wars ...

Ball-rolling bees reveal complex learning

February 23, 2017

Bumblebees can be trained to score goals using a mini-ball, revealing unprecedented learning abilities, according to scientists at Queen Mary University of London (QMUL).

New link found between sex and viruses

February 23, 2017

Sexual reproduction and viral infections actually have a lot in common. According to new research, both processes rely on a single protein that enables the seamless fusion of two cells, such as a sperm cell and egg cell, ...


Adjust slider to filter visible comments by rank

Display comments: newest first

1 / 5 (9) Jul 25, 2013
oooooold news.. popular belief is no justification
1.7 / 5 (11) Jul 25, 2013
I've been predicting this for years - most, if not all, animals, should have just under an octave of visiospatial bandwidth, just like us, regardless of the number of different types of 'cone' cells.

An interesting followup would be to discover how dogs handle the colour constancy problem - the 'relative' aspect of hue perception, that seems to directly contradict the notion of distinct, specialised colour cones tuned to specific wavelengths.

GC Huth's work here is still years ahead of the curve IMHO..
1.5 / 5 (8) Jul 25, 2013
Some dogs can see red too.. My mother had a poodle 40 years ago that hated anything red. She couldn't hang any clothes on our closeline because the dog would jump up and pull it off the line as soon as she turned her back on it and would do his best to tear it up and bury it somewhere in the back yard never to be seen again. She had to keep him in the house when the laundry was on the line. Didn't bother anything of a different color, just red stuff.

1 / 5 (3) Jul 26, 2013
Very interesting ! but semantics/linguistics do confuse!

At each instance of the article's "have two cones" , one should, rather, understand that the author must have intended to say " have two different types of cones" . Only then does the article make sense. This work suggests several topics for further research , too ...

1 / 5 (7) Jul 26, 2013
Anyone remember Land's demonstrations in the 50's and 60's that with 2 selected colors (cones) the whole rainbow of colors could be synthesized?

What did the researchers do to eliminate the possibility that the dogs SMELLED the meat then correctly chose the meat containing container? They should have inverted the experiment to eliminate this possibility.
1 / 5 (2) Jul 28, 2013
Two cones should still be able to see the whole color spectrum... one tuned to either end and gradient throughout the middle could still cover all colors. I would be surprised if they didn't see all colors.
1 / 5 (6) Jul 28, 2013
Anyone remember Land's demonstrations in the 50's and 60's that with 2 selected colors (cones) the whole rainbow of colors could be synthesized?

By far the most comprehensive followup of Land's groundbreaking work is Huth's. You can follow a link from his website to a 1980's BBC Horizon documentary covering Land's work on the colour-constancy problem - the episode's called 'Colourful Notions' - i have a copy here, it's just as fascinating 30 odd years on.

I also have a selection of Land's two-colour photos that do indeed induce a full palette of colours - only when you zoom in does it become clear you're only seeing red and white (though some plates also use black, too).

Wendy Carlos's site has a good summary of this work:


You can also find many of the two-colour images on Google's image search

While Land's original Retinex theory is now largely discredited, Huth's work picks up where Land left off, but goes much further...
1 / 5 (4) Jul 28, 2013
It's because generally, but in the fields of cognitive science especially, there's an underlying presumption of our own uniqueness and specialties, coupled with an unreasonably over-zealous stigmatisation against anthropomorphism.

Any notion that our own faculties fall within a spectrum of general abilities, rather than at the zenith, is regarded as conceptual leap. This applies as much to quantitative differences as qualitative ones...

With regards to colour perception, and the notion of a linear relationship between the number of types of colour cones and their resulting net optical bandwidth, if trichromites experience three primary colours and so forth, at what number are we to suppose that bandwidth exceeds an octave? In other words how many types of colour cones would one need to see equivalencies such as, say, an 'upper' green, beyond violet, as well as the conventional one between red and blue?

Bandwidths spanning more than one octave would present erroneous information!
1 / 5 (4) Jul 28, 2013
Nb this last point contradicts the notion that colour perception is dependent upon number of different cone types.

Rather, all complex eyes have the same, sub-octave bandwidth, transposed up or down relative to ours, depending upon their owners' optimal habitat and whether they're nocturnal, diurnal, aquatic or terrestrial etc. etc.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.