Although phenotypic plasticity is demonstrably adaptive in a range of settings, organisms are not perfectly plastic. Costs of plasticity comprise one factor predicted to counter the evolution of this adaptive strategy, yet evidence of costs is rare. Here, we test the fitness effects of plastic life-history and morphological responses to density and costs of this plasticity in recombinant inbred lines of Arabidopsis thaliana. Several costs of plasticity and homeostasis were detected. Of particular relevance, there was a significant cost of plasticity to active stem-elongation responses, an adaptive trait in many species. There was also a cost of plasticity to apical branch production at both high and low density, which resulted from the greater suppression of basal branching in genotypes with plastic apical branch production relative to genotypes with fixed apical branch production. The presence of a cost in multiple environments (i.e., a global cost) is predicted to counter the evolution of plasticity. Experimental segregating progenies such as the one used here are expected to have higher genetic costs of plasticity than arrays of genotypes sampled from natural populations because selection should remove genotypes with costs resulting from linkage disequilibrium or epistasis. The use of experimental progeny arrays therefore increases the ability to evaluate genetic costs.