We evaluated binocular summation, eye dominance, and learning in the Trivector and Ellipses procedures of the Cambridge Colour Test (CCT). Subjects (n = 36, 18-30 years old) were recruited among students and staff from the University of São Paulo. Inclusion criteria were absence of ophthalmological complaints and best-corrected Snellen VA 20/20 or better. The subjects were tested in three randomly selected eye conditions: binocular, monocular dominant eye, and nondominant eye. Results obtained in the binocular and monocular conditions did not differ statistically for thresholds measured along the protan, deutan, and tritan confusion axes (ANOVA, P > 0.05). No statistical difference was detected among discrimination ellipses obtained in binocular or monocular conditions (ANOVA, P > 0.05), suggesting absence of binocular summation or of an effect of eye dominance. Possible effects of learning were examined by comparing successive thresholds obtained in the three testing conditions. There was no evidence of improvement as a function of testing order (ANCOVA, P > 0.05). We conclude that CCT thresholds are not affected by binocularity, eye dominance, or learning. Our results differ from those found by Verriest et al. (1982) using the Farnsworth-Munsell 100 Hue test and Hovis et al. (2004) using the Farnsworth-Munsell panel D-15 test.