ABSTRACT
This chapter shows how evolution can optimize in the presence of multiple tasks. It argues that multiple tasks lead to simple geometrical patterns in biological data. These patterns can help to understand the evolutionary trade-offs at play. The chapter considers evolution toward a single objective, such as maximizing the growth rate of bacteria. The Grants lived on a tiny island in the Galapagos and observed finch evolution. Evolutionary trade-offs lead to patterns in phenotype space in which a continuum of possibilities is bounded within polyhedral-like shapes with pointy vertices. The pointy vertices can be used to infer the tasks at play. The position of each phenotype relative to the vertices tells how important each task was in its evolution. Natural selection will tend to converge to the summit of the fitness landscape, to the phenotype that maximizes fitness. Phenotypes will perhaps form a cloud around the peak due to randomizing forces.
