To elucidate the molecular mechanisms of vertebrate color vision, it is essential to establish associations between amino acid substitutions and the directions of lambda max shifts of visual pigments. In this way, we can identify critical amino acid changes that may be responsible for lambda max shifts of visual pigments. In this process, we may consider only highly conserved residues, simply because the evolutionary conservation often implies functional importance. Using such an "evolutionary model" as a convenient tool in designing mutagenesis experiments, we can test specific hypotheses on the molecular mechanisms that are responsible for color vision in vertebrates. Virtually any vertebrate opsin cDNA can be expressed in COS cells, reconstituted with 11-cis-retinal, and the lambda max values of the regenerated pigments can be measured rather easily. By constructing mutant pigments with desired amino acid changes and conducting the in vitro assay and comparing their lambda max values with those of corresponding wild-type pigments, we can elucidate the molecular mechanisms of lambda max shifts--and color vision--of vertebrates rigorously.