Recent work has revealed the molecular mechanisms governing one of the most dramatic examples of parallel evolution in nature: the repeated loss of lateral plate armor in freshwater populations of threespine stickleback. Yet, the ecological mechanisms responsible for armor loss remain unclear. Using a balanced experimental design, we examined Heuts' (1947) hypothesis that selection due to differences in salinity indirectly drive the reduction of lateral plate armor in fresh water while maintaining armor in the sea. We measured two fitness-related traits, hatching success and juvenile growth rate, in offspring of reduced (low and partial) and complete lateral plate morphs from two polymorphic populations when raised in either fresh water or salt water. In contrast to Heuts' results, there was little difference among morphs in hatching success. However, salinity strongly influenced juvenile growth: offspring of reduced lateral plate morphs grew substantially faster (up to 65%) than offspring of completely plated morphs in fresh water, but there was little difference in salt water. We suggest that the parallel loss of lateral plates in fresh water has arisen through a correlated response to selection for faster growth during lateral plate development, but the effect of salinity on hatching success and juvenile growth rate cannot explain the predominance of completely plated morphs in marine populations.