Evolutionary ecologists are fundamentally interested in how species interactions affect evolutionary change. We tested the degree to which plant-pollinator interactions affect the frequency of flower color morphs of Raphanus sativus. Petal color in R. sativus is determined by two independently assorting loci, producing four petal colors (yellow, white, pink, and bronze). We assessed the impact of pollinator discrimination on changes in flower color variation by comparing the frequency of colors produced in the presence (open pollination) versus absence (null pollination) of pollinator discrimination. We also assessed the impact of postpollination and developmental effects on progeny colors using equal pollinations with all four color morphs. Our results from open pollinations found an overrepresentation of yellow progeny in the next generation, when compared with both null pollinations and cumulative ratios based on Hardy-Weinberg and linkage equilibria assumptions. When these results were combined with those from equal pollinations, the overrepresentation of yellow could be attributed to selection from pollinators. Yet, surveys in the field the following year found no flower color frequency changes in the next generation. These results illustrate that flower color microevolution can be driven by both pollinator discrimination and other nonpollinator selective forces acting during the seed-to-adult transition, countering selection imposed by pollinators.