Trade-offs are fundamental to understanding the origin and maintenance of biological diversity. In particular, phenotypic evolution is predicted to depend critically on the shapes of the trade-off functions relating fitness costs and benefits to trait expression. Existing studies have focused on understanding the evolution of single traits. This approach assumes that traits are independent, despite the recognition that phenotypic traits are often highly integrated. This integration suggests that selection will often be multivariate, so that trade-offs underlying individual traits will affect the evolution of other phenotypic traits as well. Here I apply these insights to a number of questions. In the first chapter, I ask how selection acting on energy allocation and behavior simultaneously can lead to qualitatively different expressions of predator-induced defenses. I show that different shapes of the trade-off between foraging gain and predation risk lead to defense strategies that alter energy allocation only, behavior only, or both simultaneously in response to predation risk. The second chapter extends these insights to explore the evolutionary emergence of phenotypic polymorphism. I ask whether selection acting on both body size and behavior leads to different evolutionary outcomes from either single-trait case. I show that, depending on the shapes of the underlying phenotypic trade-offs, multivariate selection can lead to evolutionary branching when univariate selection predicts a global optimum and predicting a global optimum when univariate selection predicts evolutionary branching. The final chapter uses phylogenetic comparative analysis to study the adaptive evolution of amphibian morphology and behavior. Using measurements of trait means and plasticities, a novel phylogeny, and extensive habitat data, I test whether species differences in habitat choice have adaptive consequences for phenotypic evolution. I find strong support for the hypothesis that body size and behavioral plasticity have adaptively evolved in response to changes in pond hydroperiod and predation risk. Other morphological traits appear to be highly constrained or under strong stabilizing selection. These studies show that understanding phenotypic evolution requires considering how selection acts on the entire phenotype, rather than atomized traits, and accounting for the interactions among the trade-offs underlying trait expression.