Inferring eco-evolutionary dynamics from the imperfect fossil record of birds

The role ecology plays in deep time evolutionary dynamics, and macroevolutionary patterns, has long been disputed. Two major obstacles are our uncertainty about the ecology of extinct species, and the biased nature of the fossilization process. In this dissertation, I develop and implement a method for testing ecological reconstructions in fossil taxa by comparing predictions from standard morphometric techniques to preserved gut contents in fossil birds. These results show that certain ecological categories in birds are readily reconstructed (e.g., granivores), while others are substantially more difficult to accurate reconstruct (e.g., piscivores). I also quantify the strength and direction of fossilization bias by comparing living bird communities to recent fossil analogs (young archaeological and subfossil deposits, mass deaths, etc.). These results show that ecological bias can strongly mislead inferences about extinct ecosystems when deposit type and species-ecology are not taken into account. Finally, I combined the large morphological database I developed to explore the previous questions with a molecular phylogeny to explore deep-time dynamics in bird evolution, and document a pattern of early, fast rates of ecomorphological evolution that would be undetected without fossil data.