Computational predictions of the functional stenosis severity from coronary imaging data use an allometric scaling law to derive hyperemic blood flow (Q) from coronary arterial volume (V), Q = αV^β. Reliable estimates of α and β are essential for meaningful flow estimations. We hypothesize that the relation between Q and V depends on imaging resolution. In five canine hearts, fluorescent microspheres were injected into the left anterior descending coronary artery during maximal hyperemia. The coronary arteries of the excised heart were filled with fluorescent cast material, frozen, and processed with an imaging cryomicrotome to yield a three-dimensional representation of the coronary arterial network. The effect of limited image resolution was simulated by assessing scaling law parameters from the virtual arterial network at 11 truncation levels ranging from 50 to 1,000 μm segment radius. Mapped microsphere locations were used to derive the corresponding relative Q using a reference truncation level of 200 μm. The scaling law factor α did not change with truncation level, despite considerable intersubject variability. In contrast, the scaling law exponent β decreased from 0.79 to 0.55 with increasing truncation radius and was significantly lower for truncation radii above 500 μm vs. 50 μm ( P < 0.05). Hyperemic Q was underestimated for vessel truncation above the reference level. In conclusion, flow-crown volume relations confirmed overall power law behavior; however, this relation depends on the terminal vessel radius that can be visualized. The scaling law exponent β should therefore be adapted to the resolution of the imaging modality.