Prior to the advent of large sequence datasets, it was assumed that innovation and divergence at the morphological and physiological level would be easily explained at the molecular level. Molecular explanations for physiological adaptations have, however, been rare.