Phenotypic plasticity is predicted to evolve in environmentally variable habitats, or those experiencing a high frequency of strong selection. However, the evolution of plasticity may be constrained by costs or physiological constraints. In flowers, UV-absorbing pigmentation ameliorates UV damage to pollen, and is linked with elevated UV exposure. Whether plasticity contributes to this pattern remains unclear. Petals of Argentina anserina have larger UV-absorbing petal areas at high elevations where they experience higher and more variable UV exposure compared to low elevations. We measured UV-induced pigmentation plasticity in high- and low-elevation populations (hereafter, 'high, 'low'), and selection on pigmentation via male fitness. We dissected UV pigment biochemistry using metabolomics to explore biochemical mechanisms underlying plasticity. High displayed positive UV-induced pigmentation plasticity but low lacked plasticity. Selection favored elevated pigmentation under UV in high, supporting adaptive plasticity. In high, UV-absorption was conferred by flavonoids produced in one flavonoid pathway branch. However, in low, UV-absorption was associated with many compounds spanning many branches. Elevated plasticity was thus associated with reduced pigment diversity. Results are consistent with adaptive floral pigmentation plasticity in more extreme and variable environments. We discuss how biochemical underpinnings of pigmentation may permit or constrain the evolution of pigmentation plasticity.

Four datasets are provided in a single excel file, each on a different tab. Field_Phenotypes was the dataset used to assess elevational patterns of floral trait variation among populations across two years of field-collected data. GreenHouse_Data was the dataset used to assess whether low and high elevation populations differed in phenotypic plasticity in response to UV treatments. Poll_Germ Data was the dataset used to analyze whether selection acted differently on floral pigmentation based on UV exposure and elevation-of-origin. Flav_andTrait_Data was used to 1) generate Principle Components that described extractable flavonoid profiles of petals, 2) analyze whether variation in flavonoid PC values was explained by petal traits (flower PC1, flower PC2), elevation-of-origin, and UV treatment, and 3) correlate flower pigmentation (flower PC1) with each of hte 49 extracted flavonoid abundances.

This dataset contains phenotypic data that was collected in the field, and the greenhouse, fitness data (pollen germination rate) that was collected on plants in the greenhouse, and petal metabolomic data that was collected on plants in the greenhouse. Multiple floral traits have been distilled into PC values. 49 flavonoid compounds have been distilled into PC values. Pollen germination rate has been standardized for selection analyses.