Previous work on canine color vision has found that dogs do see colors and that colors are more important to them in many cases than the level of brightness of an object. However, studies of their eyes and the light sensitive cells in them have suggested that their color vision may be similar to people who have deuteranopia, better known as red-green color blindness.
In a recent study, scientists in Italy tested this hypothesis with a behavioral test. They showed dogs a series of colored targets that depict movement—in this case a red running cat against a green background—in a way that is not detectable unless the colors of red and green can be distinguished. The running cat was chosen because with a series of tests with dogs not used in the final experiment, a cat image in black and white prompted more reactions than the other images considered.
The reds and greens used in the test were modeled after the test for red-green color blindness in people (Ishihara’s test). Many of us are familiar with this test from school, in which green circles form the background and there is an image such as the number “26” within it that is made out of two different shades of red circles. Most people see the number easily, but to people with deuteranopia, the “6” is hidden and they report seeing only a “2”.
In this experiment, dogs were shown a series of images of the cat in each shade of red—the one that color blind people can detect, and the one that remains hidden in test images. They were also shown a series of animations using the shades of green in the background to make sure that any response was not due to the animation itself, but to the color differences in the images compared to the background. The final set of images was a control one with a black cat on a white background.
The data collected measured how much the dog oriented with head, body or eyes to the target of interest. Dogs were videotaped continuously and then scored for any of the following behavior patterns: ears up and forward, turning head from left to right, eyes wide open, forward body position, eye or ear directed toward the target, gazing at the target, head slightly lowered, paw raised, freezing, alert posture and head tilt. Also, the total amount of time that the dog spent looking at each image was recorded.
It’s a clever design because the dogs do not have to be trained to do this, allowing tests of large numbers of dogs in a short time. Additionally, it frees the experiment from any troubles of reinforcement or motivation so typical of studies requiring a learned response.
In the study, dogs spent more time looking at and had higher alerting behavioral scores to the cat that can be seen by people with red-green color blindness and the black cat on a white background than towards the cat that is camouflaged for such people or the cat that is the same green color as the background. These results support the hypothesis that dogs are red-green color blind.
This is the first test of color vision in dogs based on a modification of a test used in people, which allows a direct comparison of color blindness and color vision between the two species.