Researchers since Darwin have been intrigued by how selection on flower color and form may shape plant fitness.  Despite the extensive research on correlated selection in floral traits, researchers rarely consider that one floral trait may act as a sieve on the selection regime of other floral traits. We therefore ask the following: can the presence or absence of flower color modulate selection on other floral traits? To test this question, we conducted a common-garden experiment with sixth-generation offspring from a cross of two sister species of morning glory: Ipomoea lacunosa and I. cordatotriloba. Each line had a mixed genetic background that was scored in the greenhouse for the presence or absence of flower color and measured for several flower size and sugar concentration traits (see Liao et al 2021). In the field, all plants from four replicate plots of fifty lines were weighed to get an estimate of seed set per line. To test if selection on floral traits differs for pink rather than white limbed lines, we ran regressions of seed flower size, sugar concentration, and nectar concentration on seed production via total seed mass and tested if these predictions changed when considering pink vs. white limbed lines. Corolla length, the representative size trait selected, and sugar concentration were respectively suggestive (p-value=0.0875) and significant (p-value=0.0488) predictors of seed set. However, when accounting for flower color, only pink-limbed lines showed a positive relationship between these traits and seed set, with a statistically suggestive effect of color on sugar concentration and flower size (p-values=0.0850, 0.0637, respectively) via ANCOVA. These results suggest that directional selection for increased flower size and sugar concentration only occurs for pink plants and perhaps provide evidence that color may act as a selective sieve for other floral traits.