A long-standing question in evolutionary biology asks whether the genetic changes contributing to phenotypic evolution are predictable. Here, we identify a genetic change associated with segregating variation in flower color within a population of Mimulus lewisii. To determine whether these types of changes are predictable, we combined this information with data from other species to investigate whether the spectrum of mutations affecting flower color transitions differs based on the evolutionary time-scale since divergence. We used classic genetic techniques, along with gene expression and population genetic approaches, to identify the putative, loss-of-function mutation that generates rare, white flowers instead of the common, pink color in M. lewisii. We found that a frameshift mutation in an anthocyanin pathway gene is responsible for the white-flowered polymorphism found in this population of M. lewisii. Comparison of our results with data from other species reveals a broader spectrum of flower color mutations segregating within populations relative to those that fix between populations. These results suggest that the genetic basis of fixed differences in flower color may be predictable, but that for segregating variation is not.