A Research Proposal:

Color Perception Capacity ofDogs

Meagan L. Anderson

Stephen F. Austin StateUniversity

The planet Earth is covered with many colors, bothnatural and man-made. The human perception of color is morefrequently discussed than the actual visible radiation. This may bedue to the emotional link that humans have to their perceptions ofcolors (Color and Acuity, 1998). Artists who use certain colors todepict particular emotions, and the new field of study, which hasrecently developed, connecting colors to emotional meanings both showthis link. Most everyone has a favorite color with which theyidentify themselves. Color is therefore not only a prominent part ofhuman vision, but is an important part of emotion.

Although some animals are not capable of colorperception (as found through scientific study), the ability todiscriminate between colors serves an important function to those whoare (Williams & Hogan, 1994). Humans use color as a means toseparate and organize the environment (Goldstein, 2002). For example,when cooking meat, the percentage of red or pink in the center allowsthe chef to determine when it has been cooked thoroughly. Also, whenshopping for fresh fruit, the color sends a signal as to thereadiness of the pieces. Even picking out a white bloom on a greenplant requires color vision. Some animals must rely on their abilityto discriminate colors in order to hunt, spot predators, or gatherfruit. They may also use color to indicate sexual availability, whichaides in reproduction (Color Experience, 2001). Though humans mayrely on color for certain activities, other animals use it to escapebeing captured by prey, and perpetuate life through reproduction.This is why evolution is thought to be responsible for color vision.The animals with color vision were better suited to survive, as wellas seek out mates, and pass on their genes.

According to Hartmann, Dobson, and Hainline(2000), about five percent of humans do not have normal color vision.Females are much less likely to have any type of color-deficiency dueto its genetic properties. Color deficiency in humans is theinability to perceive some of the colors in the normal spectrum ofvision for humans (Goldstein, 2002). People who have acolor-deficiency have the ability to adapt to their environment, muchthe same way other animals are expected to adapt, although certaintasks, such as picking out objects against a textured background(with one eye only), will always be more difficult (Corn & Webne,2001).

The most common form of color-deficiency isred-green deficiency, which is often indicated by the inability todiscriminate between the two (Effective Color Contrast, 2002). Thetwo forms of red-green deficiencies are protanopia and deuteranopia(Goldstein, 2002). For people with deficient color vision, particularcolor schemes are easier to look at and use (Merchant, 2000). Anexample of a particularly difficult set-up is a green object on a redbackground. This color scheme is thought to be difficult for humansas well as dogs.

Popular belief about the color vision of dogs wasthat they had none, and were therefore colorblind (What Do Dogs,1998). Total colorblindness indicates that their entire world wouldbe shades of gray. Now, many researchers believe that dogs are ableto see very similarly to an average human deuteranope (Color andAcuity, 1998). It is therefore assumed that dogs also lack the redcone type. Dogs have excellent night vision, and therefore they havea large number of rods. In order to have room for so many rods, theyhave fewer cones, which sacrifices color vision. This allows them tobe better adapted for their lifestyles. This also backs up the theorythat evolution caused color vision or a lack of it. When looked at,dogs' central retinas only contain twenty-percent cones, leaving theother eighty-percent to be rods (used for dim lighting). The centralretinas, or fovea, of humans are made up entirely of cones. Thiswould explain why humans can see colors better, but lack the nightvision that dogs possess (Performance Descriptors, 2000).

Many tests have been run in order to determinewhat colors dogs can see, but no information was found on the abilityto differentiate between shades of the same color. In the followingexperiment, the dogs' ability to discriminate colors will be moreclosely tested. It is expected that the dogs will be able to choosemore accurately between shades of blue and yellow than shades of redand green.



All of the participants will be randomly selecteddogs of various breeds and combinations of breeds from no-killadoption agencies. There will be both male and female dogs of varyingages. They will receive extra treats, play time, and affection aftereach testing session, regardless of ability to cooperate with theexperimenter. No dog will be withheld any of the former due to aninability to participate.


There will be nine 12" by 12" squares ofcardboard. Three of them will be white, three will be gray, and theremaining three will be black. Also, twenty-four small fish bowlswill contain water with food coloring of different concentrations (tocreate different shades of colors). In previous research, manydifferent stimuli have been used, but these will be used because ofavailability. Dog treats will be used in order to reinforce correctbehaviors during training. Tennis balls will also be used asreinforcement, by allowing the dogs to play with them at the end ofthe trials.


Each dog will undergo preliminary training withthe experimenter(s), in which they will learn to touch theodd-colored stimuli with their noses. The initial stimuli will be thecardboard squares. They will be presented in groups of three, withtwo of them matching in color. The odd-colored square must not be putin the same place every time. A dog treat will be placed under thesquare that does not match. Upon nosing the correct square, the dogwill receive the treat. Once the dog begins to make the correctselection on the first time for each trial, the experimenter shouldhold the treats and hand feed them to the dog for further correctchoices. If the dog fails to make the transition to the new form ofreinforcement, the training will start over. On the other hand, ifthe dog makes the change well, then the training phase may be ended.

The testing phase will begin by using the waterstimuli, which will consist of two bowls with identicalconcentrations of coloring, and one of a different concentration.This will create a setup in which three stimuli will be presented,with one being different. The dogs will be tested for shades of blue,green, yellow, and red. Records should be kept for which colorcombinations they accurately discriminate as well as those they donot. Each color combination should be tested, and treats should begiven after each trial in which a correct response ismade.


This will be a two-factor within-subjects design.The levels of the first independent variable will be the colors(blue, red, yellow, and green). The second IV will be the shade ofthe colors with four levels (5, 10, 15, or 20 drops of fodd coloringin 8 oz. of water). The dependent variable will be the ability todetect the stimuli that differ in color. The ability to discriminatewill count as a correct answer, while the inability will be shown asan incorrect answer.


The data that will be collected will be based on acorrect or incorrect response for each trial. The results of thewithin-subjects experiment should support the hypothesis that dogscan discriminate between shades of blue and yellow with much betteraccuracy than they can with red and green. The results will besubmitted to a two- factor within-subjects ANOVA. A main effect forcolor is expected, with blue and yellow being easier to discriminate.No main effect is expected for shade. An interaction effect isexpected in that the darker shades of the yello w and blue lead toslightly better accuracy. The color(s) which has(have) the highestrate of correct responses will be assumed to be easier for dogs todiscriminate between.


The results of this experiment should provide aclear evaluation of dogs' ability to discriminate between shades ofcolors. This experiment should also have results that are consistentwith the color spectrum shown on Dr. P's Dog Training (1998). Becausestudies such as the ones on the Dr. P's web site have found that dogssee in shades of blue, yellow, grey, and white, it would bereasonable to assume that they could therefore discriminate betweenshades of these colors much better than other colors.

The hypothesis and experiment assumes that thedogs have normal vision for canine standards. It is however,important to remember that dogs have vision problems unique to theindividual. For example, rod-cone degeneration occurs in somegenetically predetermined dogs. Kommonen, Kylma, Karhunen, Dawson,and Penn (1997) state that this impairs the visual acuity as well ascolor perception. This type of problem would be a factor in skewedresults, but due to the participant pool, these types of problems arenot tested for.

The results of this experiment should lead tofurther experimentation into the shades of blue and yellow that aremost easily seen or discriminated. This research could benefit dogowners, veterinarians, as well as breeders.

Works Cited

Color and Acuity Differences between Dogs andHumans. (1998). Dr. P's Dog Training: Vision in Dogs & People.Retrieved March 2, 2002 from: http://www.uwsp.edu/psych/dog/LA/davis2.htm

Color Experience and the Human Animal. (2001). AICColor '01 Rochester. Retrieved March 3, 2002 from: http://www.iscc.org/aic2001/abstracts/symposium/What_is_Color/Hardin.doc

Corn,E., & Webne,S. (2001). Visual impairmentsand what they mean for participation in activities. Journal of VisualImpairments & Blindness, 95(2), 110-118.

Effective Color Contrast: Designing for Peoplewith Partial Sight and Color Deficiencies. (2002). LighthouseInternational. Retrieved March 4, 2002 from: http://www.lighthouse.org/color_contrast.htm.

Goldstein, E. Bruce. (2002). Sensation andPerception. Pacific Grove, California: Wadsworth.

Hartmann, E., Dobson, V., & Hainline, L.(2000). Preschool vision screening: summary of a task force report.Pediatrics, 106, 1105-1112.

Kommonen, B., Kylma, T., Karhunen, U., Dawson, W.,& Penn, J. (1997). Impaired retinal function in young Labradorretriever dogs heterozygous for late onset rod-cone degeneration.Vision Research, 37 (3), 365-370.

Merchant, D. (2000). The trouble with color.Library Computing: Internet & Software Applications forInformation Professionals, 19 (3/4), 208-212.

Performance Descriptors Based on Processes inAnimal Vision. (2000). Color Vision. Retrieved March 2, 2002 from:http://www.4colorvision.com/perform.htm

What Do Dogs and Cats See? (1998). VeterinaryVision: Animal Eye Specialists. Retrieved from: http://www.veterinaryvision.com/See/htm

Williams, R., & Hogan, D. (1994). Targetrecognition and visual maps in the thalamus of achiasmatic dogs.Nature, 367 (6464), 637-639.