Date:
Title: The evolution and development of the insect egg and ovary
Abstract: Life comes in many shapes and sizes. Efforts to describe, categorize, and understand this diversity have been a driving force behind biological discovery. Evolutionary developmental biology, or evo-devo, seeks to understand its origins by investigating how evolutionary changes to the developmental process can lead to new traits. But the field of evo-devo is at an inflection point. In recent decades, rich and complex datasets have become more accessible and inexpensive to generate from a wide variety of organisms. The outcome is that we have unprecedented opportunities to investigate the developmental basis of diversity. However, our ability to use these data to make robust inference will require us to compare traits across taxa using statistical methods rooted in evolutionary theory.
In this thesis, I investigate the origins of shape and size diversity through the application of evolutionary comparative methods to large datasets of morphological and developmental traits. In Chapter 1, I describe the state of the field of evo-devo, and advocate for a shift in the framework often used in analyses, toward one rooted in evolutionary tree-based thinking. Chapter 2 presents a dataset of 10,000 descriptions of insect egg size and shape, assembled using custom software tools to extract descriptions from the published literature. Chapter 3 analyzes this dataset on a phylogenetic tree of insects to test hypotheses about size and shape evolution in relation to ecological and developmental features. Chapter 4 combines egg size data with a dataset of more than 3,000 ovary descriptions to test a longstanding hypothesis about the size and number of offspring. Chapter 5 investigates the evolutionary relationships of a specific lineage of Hawaiian flies in the family Drosophilidae, identifying specific evolutionary shifts in egg and ovary diversification.
Throughout this thesis I explore the relatively untapped resource of "big data" that is the published literature. Carefully catalogued and preserved in biological libraries are many thousands of biological descriptions, drawings, photographs, and hypotheses. New efforts to digitize these works have helped unlock their potential. Here I present several studies that use large datasets assembled from the published literature to revisit old and widely-accepted hypotheses about insects. My findings show that many of these hypotheses have limited utility in explaining the variation we see across large datasets. Instead, as we compare data on biological variation in the context of evolutionary trees, we see that explanations that might have served us well in one group of organisms are unlikely to represent generalizable rules.
Committee: Cassandra Extavour (Advisor), Casey Dunn (Yale), Gonzalo Giribet, Mansi Srivastava