Why are some fruits toxic? glycoalkaloids in Solanum and fruit choice by vertebrates

We examined the influence of secondary metabolites (glycoalkaloids) of Solanum fruits on fruit consumption by seed dispersers and seed predators. Our goal was to determine the degree to which secondary metabolites might explain fruit-frugivore interactions that heretofore have not been explained by other fruit-related variables (e.g., color, nutrient content, seediness, etc.). Using feeding trails with real fruits and artificial fruit pulp media, we tested two sets of hypotheses, one relating to the specificity of secondary metabolites and the other to intra- and interspecific variation in secondary metabolite content. First, the Directed Toxicity Hypothesis was tested against its alternative, the General Toxicity Hypothesis. The Directed Toxicity Hypothesis posits that fruit pulp secondary metabolites are directed primarily toward organisms that destroy seeds following fruit consumption, remaining nontoxic to organisms that disperse seeds. The General Toxicity Hypothesis states that these secondary metabolites are general in their effects. Second, the Removal Rate Hypothesis was tested against its alternative, the Nutrient-Toxin Titration Hypothesis. The Removal Rate Hypothesis posits that profitable fruits, with high natural rates of removal, contain only low levels of secondary metabolites as defenses against seed predators and pathogens. In contrast, the Nutrient-Toxin Titration Hypothesis states that profitable fruits are better able to afford protection by high levels of secondary metabolites because high nutrient levels compensate for deterrent effects of secondary metabolites. We compared Solanum carolinense, with high levels of secondary metabolites (glycoalkaloids) in ripe fruit, and S. americanum, with negligible levels, conducting four sets of feeding trials: (1) Real Fruit trials to determine the preferences of our study animals for real fruit of each plant species; (2) Fruit Mimic trials using artificial fruit pulp media to determine the relative effects of nutrient and glycoalkaloid content on fruit preferences; (3) Interaction trials using artificial fruit pulp media to examine relative and interactive effects of the two main glycoalkaloids in our fruits (α-solasonine and α-solamargine); and (4) Nutrient-Toxin Titration trials using artificial fruit pulp media to determine how much nutrient content might compensate for deterrent effects. Both seed predators and seed dispersers preferred the low-glycoalkaloid S. americanum fruits and were de