Genomic parallelism defines repeated evolution of an inducible offense.

A common developmental response to resource competition is an inducible offense, the facultative predation of competitors. At its extreme, this response involves the development of alternative phenotypic morphs, or polyphenism. However, how polyphenism evolves to meet ecological challenges, such as competitor species, is unknown. Using replicated experimental evolution, during which starved nematodes could consume heterospecific competitors, we investigated whether induction of a predatory morph could evolve and how generalizable this change's genetic basis is. Fifty generations of evolution across multiple populations resulted in parallel changes in higher morph-induction and parallel genomic responses, including repeated selection for a specific transcription-factor binding-site variant. In tandem, we artificially selected directly for tooth morphology and drove the predatory morph near to fixation. That trait-specific selection promoted greater changes in predatory morph induction than experimental evolution indicates that polyphenism evolution is balanced by selection for whole-organism performance. Our results thus describe the predictability by which a resource polyphenism evolves amid scarce resources.