The resolution to the color-metallicity problem is the use of multi-metallicity models. These models use an idealized metallicity distribution (from Kodama & Arimoto) where each model is parameterized by a linear compression of the metallicity distribution, fm.

Operationally, the smooth metal distribution is broken into six metallicity bins ([Fe/H]=-2.2,-1.7,-1,-0.4,0,0.4; Bruzual & Charlot SED models) and summed to produce total colors. The fraction of 1.0 is set to reproduce the colors of an L* elliptical. To reproduce a sequence of lower metallicity galaxies, the metal-poor end is held fixed and the peak is lower by constant fractions to reproduce lower metallicity galaxies.