Information about the time-dependent production and temperature-dependent diffusion of radiogenic argon in feldspars can be used to constrain the thermal evolution of meteorites, mountain belts, intrusive magmatic bodies, and a host of other Earth and planetary processes. To better asses the accuracy of such thermal models, an understanding of the mechanisms, pathways, and processes by which argon diffuses in feldspars is required. In this thesis I present step-heating diffusion experiments conducted on feldspars with diverse compositions, structural states, and microstructural characteristics. The experiments reveal systematic variations in diffusive behavior that appear closely related to these variables, with apparent closure temperatures for 0.1-1 millimeter grains of ~200 to 400 °C (assuming a 10 °C/Ma cooling rate). Given such variability, there is no broadly applicable set of diffusion parameters that can be utilized in feldspar thermal modeling; sample-specific data are required. Diffusion experiments conducted on oriented cleavage flakes do not reveal directionally-dependent diffusive anisotropy to within the precision limits of our approach. Additional experiments aimed at constraining the physical significance of the diffusion domain are presented and indicate that unaltered feldspar crystals with or without coherent exsolution lamellae diffuse at the grain scale, whereas feldspars containing hydrothermal alteration and/or incoherent sub-grain intergrowths do not. Arrhenius plots for argon diffusion in feldspars appear to reflect a confluence of intrinsic diffusion kinetics and structural transitions that occur during incremental heating experiments. These structural transitions, along with sub-grain domain size variations, cause deviations from linearity (i.e., upward and downward curvature) on Arrhenius plots. Detailed descriptions of the structural responses of different feldspars to heating are given, and an atomistic model for coincident Arrhenius behavior is proposed. The resulting implications for accurately extrapolating laboratory-derived diffusion parameters to natural settings and over geologic time are discussed. The data indicate that considerable inaccuracies may exist in published thermal histories obtained using multiple diffusion domain (MDD) models fit to exsolved alkali feldspar Arrhenius plots, where the inferred Ar partial retention zones may be anomalously hot.