It is generally thought that random mutations will, on average, reduce an organism's fitness because resulting phenotypic changes are likely to be maladaptive. This relationship leads to the prediction that mutations that alter more phenotypic traits, that is, are more pleiotropic, will impose larger fitness costs than mutations that affect fewer traits. Here we present a systems approach to test this expectation. Previous studies have independently estimated fitness and morphological effects of deleting all nonessential genes in Saccharomyces cerevisiae. Using datasets generated by these studies, we examined the relationship between the pleiotropic effect of each deletion mutation, measured as the number of morphological traits differing from the parental strain, and its effect on fitness. Pleiotropy explained ∼18% of variation in fitness among the mutants even once we controlled for correlations between morphological traits. This relationship was robust to consideration of other explanatory factors, including the number of protein–protein interactions and the network position of the deleted genes. These results are consistent with pleiotropy having a direct role in affecting fitness.