Plate tectonic models succeed at predicting observations at scales of 100--1000 km, but tend to fail for observations at outcrop scales. Such predictions become nebulous when considering small-scale heterogeneous structure originally generated in the lower crust or upper mantle, whether the observations are made from outcrop, from boreholes, or remotely by seismic methods. Earth scientists have classically dealt with the problem of small-scale heterogeneity by treating it as noise from both an observational and modeling standpoint; however, the increased comprehension of the scale invariance of many geologic processes has spurred investigators to attempt to characterize observations at all available scales. Modelers have responded by adopting more of a statistical characterization approach for spatial variability. Heterogeneity in the Crust and Upper Mantle offers a state-of-the-art assessment of how the difficulty of understanding small-scale heterogeneity is being addressed by active researchers across a spectrum of disciplines, including structural geology, seismology, and statistical physics.