Dogs see the world in black and white.

For most of the twentieth century, it was taught as settled fact: dogs live in a world drained of color, perceiving everything in shades of grey, black, and white. The claim appeared in veterinary guides, popular science books, and classrooms from the 1940s onward. It seemed plausible, dogs' brains and sensory equipment were clearly different from humans', and it was easy to assume that a simpler nervous system might process vision more crudely.

The assumption was wrong, and the evidence against it accumulated slowly over decades before a definitive study resolved the question.

The belief appears to have its roots in nineteenth and early twentieth century assumptions about animal cognition. Behaviorists argued that attributing complex perception to animals was anthropomorphism; the safer scientific position was to assume a minimal inner life until proven otherwise. This conservatism, reasonable in its origins, calcified into a specific factual claim that outlived its evidentiary basis.

Researchers had been testing animal color vision since the 1930s and 1940s using conditioning experiments: animals trained to respond to a particular color and then tested on different wavelengths. The results for dogs were mixed and methodologically contentious. Some early experiments suggested no color discrimination; others suggested limited discrimination but could not rule out brightness cues as the real variable. The question remained genuinely open, but the textbook version treated it as closed, in favor of the no-color answer.

The resolution came from the lab of Gerald Jacobs at the University of California, Santa Barbara, which had spent decades mapping the visual systems of mammals. In 1989, Jacobs, Jay Neitz, and Timothy Geist published 'Color Vision in the Dog' in Visual Neuroscience. Using carefully controlled psychophysical testing, they trained dogs to discriminate between lights of different wavelengths on a task designed to eliminate brightness as a confound. The results were unambiguous: dogs possessed meaningful color discrimination, but it was limited to a range comparable to a human who is red-green colorblind.

The physiological explanation lay in the photoreceptors. Dogs have two types of cone cells, rather than the three that give most humans trichromatic vision. Electroretinographic measurements placed these cones' peak sensitivities near 429 nanometres, in the blue-violet range, and 555 nanometres, in the yellow-green range. With these two types of cones, a dog can distinguish blue from yellow and can perceive varying saturations across that axis of color space. What falls outside their discrimination are the distinctions that require a third cone type: the difference between red and green, which both appear as a similar yellowish or brownish shade to a dichromatic eye.

This means the world a dog inhabits is not the black-and-white of an old film but something closer to the view of a person with deuteranopia, a form of red-green colorblindness affecting roughly six percent of human men. Colors are present, but the palette is narrowed. A red ball thrown on green grass presents a genuine visual challenge to a dog: the ball and the grass may appear similarly colored, which is why dogs tracking thrown objects in fields sometimes overshoot and circle back. A yellow ball on the same grass is far easier to follow.

The teaching error persisted for decades partly because the word colorblind in everyday language implies total absence of color, which translated into the simplified textbook claim. It persisted also because nobody had conducted the right experiment carefully enough to overturn the assumption until Neitz, Geist, and Jacobs did so in 1989. The scientific literature moved on; many classrooms and popular references did not.

Dogs do not see a grey world. They see a world of blues, yellows, and the tones between, not the human palette, but a palette nonetheless.