Phenotypic and genetic clines frequently evolve due to varying selection along environmental gradients. However, the specific environmental factors that impose differential selection are often multivariate and difficult to tease apart. We addressed this complexity using a factorial manipulation of watering, soil nutrients, and (simulated) herbivory in controlled conditions to better understand the agents of selection driving the evolution of clines in a polymorphic chemical antiherbivore defence, hydrogen cyanide (HCN), of the plant white clover (Trifolium repens L.). We found the presence or absence of the two metabolic components required for HCN production (cyanogenic glucosides and linamarase) to be more prominent determinants of vegetative growth and sexual reproduction in T. repens than HCN itself. We also found that the ability to produce one or both of cyanogenic glucosides or linamarase resulted in a growth advantage under drought and simulated herbivory that outweighs the metabolic cost of their production. These results support the view that the metabolic components underlying HCN play important roles beyond defence by increasing plant tolerance to stress. The growth advantage under drought, however, was diminished in the absence of nutrient addition, consistent with multivariate interactions as drivers of selection. This study provides novel insight into how a cosmopolitan plant has adapted to environmental gradients, and more generally, highlights the importance of considering interactions between multiple environmental factors when studying the evolution of phenotypic and genetic clines.

This dataset was collected through a greenhouse experiment conducted from October 2019-January 2020. The purpose of this experiment was to assess the impact of soil moisture, soil nutrients, and herbivory as drivers of selection at two loci involved in chemical defense production in white clover (Trifolium repens). Explanatory variables include: Moisture, Nutrients, Herbivory (simulated), and Cyanotype (refers to the hydrogen cyanide plant defence phenotype). Response variables include: Whether or not plants flowered, Whether or not plants produced seeds, Number of flower heads, Seed set mass, Plant dry biomass, and Average percentage of leaf surface area damaged by herbivores.

R/RStudio and Excel (or open-source equivalent, such as Google Sheets) are required to open and use these data files.