The dependence of scale in estimates of permeability or transmissivity in pumping tests is addressed in terms of internal architecture. The need for using geologic models that go beyond treating heterogeneity as a multinormal logarithmic field is noted. By using stratification as a vehicle and simulating flow, the role of the internal architecture of the aquifer is demonstrated. Specifically, transient data generated with the help of a stratified model are evaluated without consideration of internal architecture or geological imprint, and the consequences on estimates of aquifer properties are examined particularly in terms of scale. Evidence of scale effects should suggest the need for evaluating pumping tests with models that incorporate geologic considerations.