In this section, we use our measurements to evaluate several different ideas concerning pattern and color appearance. We begin by testing a pattern-dependent linear model. This model tests the two main linearity assumptions of contrast-homogeneity and pattern-superposition. According to this model, each spatial pattern is transformed to a new color coordinate frame. The color coordinate frame for each pattern is free to depend upon the pattern. The matches to mixtures are predicted by the sum of these pattern-dependent representations.