Connecting the selective forces that drive the evolution of phenotypes to their underlying genotypes is key to understanding adaptation, but such connections are rarely tested experimentally. Threespine stickleback (Gasterosteus aculeatus) are a powerful model for such tests because genotypes that underlie putatively adaptive traits have been identified. For example, a regulatory mutation in the Ectodysplasin (Eda) gene causes a reduction in the number of bony armor plates, which occurs rapidly and repeatedly when marine sticklebacks invade freshwater. However, the source of selection on plate loss in freshwater is unknown. Here, we tested whether dietary reduction of phosphorus can account for selection on plate loss due to a growth advantage of low‐plated fish in freshwater. We crossed marine fish heterozygous for the 16 kilobase freshwater Eda haplotype and compared the growth of offspring with different genotypes under contrasting levels of dietary phosphorus in both saltwater and freshwater. Eda genotype was not associated with growth differences in any treatment, or with mechanisms that could mitigate the impacts of phosphorus limitation, such as differential phosphorus deposition, phosphorus excretion, or intestine length. This study highlights the importance of experimentally testing the putative selective forces acting on phenotypes and their underlying genotypes in the wild.