Patterns of biodiversity emerge as a product of numerous drivers including species sorting along modern and ancient environmental gradients, long-term climate changes, speciation and extinction events, the formation and weathering of mountain ranges, and plate tectonics. The fossil record is a natural experiment that records faunal change under myriad long-term processes. My dissertation focuses on changes in North American mammal community structure under conditions of long-term climate change during the Cenozoic. I ask three primary questions: Are patterns of biodiversity sensitive to the processes of fossilization? When did the modern latitudinal diversity gradient emerge and were changes driven by climate? How has community composition changed as rates of extinction fluctuated under climate change? How can we best search for the proximate climatic drivers (i.e. changes in spatial climate patterns) of diversity changes? I show, through the development of a novel simulated fossilization method, that measuring latitudinal richness gradients in the fossil record is problematic when rates of species loss are high. However, estimating the magnitude of the latitudinal diversity gradient is improved with the use of β diversity as a metric (so called latitudinal turnover gradients). I also show that the late Cenozoic (36 Ma – present) North American mammal diversity gradient varied but neared modern magnitudes only during the late Miocene and late Pleistocene. From ~15 Ma – 0.01 Ma, hoofed mammals show a decline in relatedness of species in North American communities, reflecting the phylogenetically dispersed extinction of dicot specialists as the global climate cooled. I make direct comparisons of a-historical and historical approaches to studying biodiversity change and suggest that understanding the contribution of macroevolution is requisite to testing hypotheses on the formation of modern communities. I also address the need for a high resolution terrestrial paleoclimate record and show that stable oxygen isotopes from pronghorn (Antilocapra americana) enamel are excellent sources of data on seasonal changes in environmental waters. My dissertation is a critical step in answering “What factors have led to the emergence of communities as we know them today?” and understanding the role of climate change in shaping modern and ancient communities.