Print

Calsbeek, Brittny

Biology

2010

PhD

A seminal challenge in evolutionary biology is to determine how natural selection on phenotypic variation translates into the evolution of a population overtime. Complex interactions between individuals, their genetic architecture, and the environment make it difficult to isolate the mechanisms that shape the evolutionary trajectory of a population. Consequently, much of what we currently understand about evolution comes from field and laboratory studies that focus on subsets of tractable phenotypic variation (e.g., discrete polymorphism). Evolutionary trends are predicted in many of these studies, using principles of quantitative genetics.
For example, the multivariate breeder's equation (R = GP⁻¹S) is central to both empirical and theoretical studies of how the mean phenotype of a population changes over time. Other approaches include comparing the strength and form of selection over time, and visualizing individual episodes of selection as a surface with peaks and valleys in fitness. Each ofthese empirical, theoretical and statistical approaches has its associated strengths and weaknesses, but together they form a firm foundation for understanding evolution in natural populations. In this thesis, I explore each of the fundamental approaches to studying evolution, providing evidence for multivariate selection in nature, a quantitative genetic application of the breeder's equation, and new statistical methods for analyzing multivariate selection.