Kīlauea Volcano is an outstanding natural laboratory in which to investigate basaltic volcanism, owing to over a century of extensive study. In the last decade, there has been a paradigm shift in our understanding of volcanic plumbing systems, moving away from a model where large vats of melt are stored in the crust, to a model dominated by extensive piles of settled crystals. This thesis examines the petrological evidence for this ‘mush-dominated’ model at Kīlauea Volcano, using a variety of state-of-the-art microanalytical techniques. Chapters 3 and 4 demonstrate the power of electron backscatter diffraction (EBSD) to investigate the processes leading to the formation of crystal aggregates and deformed crystals, which are observed in a wide variety of basaltic systems worldwide. Chapter 4 applies microstructural techniques developed through the study of experimentally- and naturally-deformed mantle peridotites to erupted volcanic crystals, demonstrating that mush piles have vertical extents of ∼200–700 m. Chapter 5 shows that relatively primitive Kīlauean olivines and their melt inclusions are out of major and trace element equilibrium with their carrier melts, with broad-reaching implications for the interpretation of melt inclusion records in mush-rich volcanic systems. Specifically, this work demonstrates that inclusions record processes happening over centuries to millennia, rather than those specific to an individual eruption (e.g., volatile fluxes, triggering mechanisms). Chapter 6 capitalizes on recent advances in LA-ICP-MS, permitting analyses of chalcophile elements within melt inclusions and matrix glasses. This dataset provides new constraints on the behaviour of these poorly-understood elements during mantle melting, crustal storage and fractionation, and low pressure degassing. Chapter 7 quantifies the CO2 budget of melt inclusions from the 2018 eruption using SIMS and Raman Spectroscopy. These paired analyses, combined with detailed evaluation of different CO2-H2O solubility models, yield storage depths which align remarkably well with reservoir depths inferred from geophysical inversions. Overall, this thesis demonstrates that the integration of different data streams obtained from state-of-the-art microanalysis has the potential to further our understanding of magmatic processes, even at extremely well-studied volcanic systems.