Flowers have long been considered one of the hallmarks of angiosperm evolution. They are morphologically complex structures that both promote efficient pollination and protect the developing embryo. When it was championed in 1793 by Christian Konrad Sprengel, this view of the role of flowers in reproduction, however, was highly controversial: how could a form so beautiful and pure as a flower ever be involved in something as vulgar as reproduction? Sprengel and his predecessor, Josef Köhlreuter, are considered the founders of pollination biology, and their work set the stage for that of Charles Darwin nearly a century later. Darwin saw the interaction between flowers and their pollinators as a prime example of the power of natural selection. This approach to studying the evolution of flowers-of focusing on the biotic drivers of floral morphological change-has dominated our understanding and interpretation of floral evolution. Yet, new evidence suggests that extrinsic, abiotic factors and the costs of producing and maintaining flowers may also have influenced the evolution of floral form. These non-pollinator agents of selection could represent another major shift in our understanding of how flowers have evolved. The series of studies presented in this dissertation takes one important resource, water, and examines how the requirements of providing water to flowers may influence their functioning and evolution. Two complementary approaches are used in these studies: (1) physiological measurements of the dynamics of water use on a few species and (2) comparisons of hydraulic traits for diverse sets of species. Together, these two approaches show the variability of flower water use, the anatomical traits associated with the flux of water through flowers, and how these physiological traits-and, by extension, the water requirements of flowers-vary among extant species. Together, these studies support the conclusion that maintaining flower water balance has been an important factor influencing floral evolution and, more generally, angiosperm ecology. Three studies are presented that seek to measure, using different approaches, how the water flux to flowers and the hydraulic efficiency of flowers varies among species (Chapters 1-3), within species throughout floral development (Chapters 1 and 2), and diurnally with changing environmental conditions (Chapter 3). Using a new implementation of the heat ratio method for measuring sap flow (Chapter 1), I found that sap flow velocities to flowers and inflorescences vary diurnally, throughout floral development, and among species and microhabitats. Such high variability suggested that a better approach to comparing the hydraulic architecture of flowers would be to measure the maximum efficiency of the floral hydraulic system. In Chapter 2, I quantified for a phylogenetically diverse set of species the maximum hydraulic conductance of whole flowers. This, too, was highly variable among species, as were other hydraulic traits, and the variation in all traits was driven by just two genera of early-divergent angiosperm lineages. Variation in these traits highlighted the existence of two seemingly discrete hydraulic strategies: one strategy is to maintain a high hydraulic conductance and continuously import water via the xylem while the other strategy is to have a low hydraulic conductance with long water turnover times, slow desiccation rates, and presumably high hydraulic capacitance. Investigating the tradeoffs among these strategies further, Chapter 3 focused on characterizing the water relations of flowers of two Calycanthus species, which had among the highest hydraulic conductances measured in Chapter 2. Consistent with my predictions, high hydraulic capacitance in flowers mitigates the reliance on continuous xylem delivery of water. As a result, despite maintaining a high maximum hydraulic conductance (Chapter 2), Calycanthus flowers hydraulically underperform most of the time, reaching their maximum hydraulic conductance only when turgor loss is already inevitable. The results from Chapters 2 and 3 together suggest that the monocots and eudicots, compared to the ANITA grade and magnoliids, developed thicker cuticles and reduced their stomatal abundances, which together reduce rates of water loss from flowers and prolonged the time that these flowers can remain turgid without the import of new water. Having characterized in Chapter 2 some of the anatomical traits that correlate with the hydraulic capacity of flowers, I sought in Chapters 4 and 5 to examine for a large set of species how these traits have evolved and vary among species. Specifically, I asked three questions: (1) Has there been coordinated evolution of water balance traits within flowers, which would suggest that maintaining water balance has been an important component in floral evolution? (2) Is there modularity in hydraulic trait evolution, such that flower and leaf traits have evolved independently? (3) Have hydraulic traits been under natural selection? The results from these two chapters strongly support the conclusions that floral hydraulic traits are under selection, that maintaining water balance has been an important component of floral trait evolution, and that hydraulic traits have evolved independently in flowers and leaves. These results show, for the first time, the importance of water balance in floral evolution and highlight that the physiological demands of and constraints on flowers may provide a strong counterbalance to selection by animal pollinators. As yet, studies of the physiology of flowers have received little attention and have been ignored in physiological trait databases. As a result, there has been no overarching theory describing or predicting patterns of variance in floral physiological traits. This series of studies is a first attempt at providing such a framework for predicting how floral physiological traits may vary among species and how this may differ between reproductive and vegetative traits. Although it focuses only on traits associated with the movement of water, the results show that there may be consistent trait associations and syndromes among flowers, regardless of morphology. This should be a first step in understanding how flowers function physiologically and how their functioning may vary with a variety of ecological factors and over evolutionary timescales.