Within-layer regional myocardial flows in the left and right ventricles (LV, RV) and in LV with increased myocardial workload (beta(1)-adrenoceptor stimulation) were studied transmurally in anesthetized rabbits. Myocardial flow distribution was visualized with resolutions between 0.1 x 0.1- and 1 x 1-mm(2) pixels, using digital radiography combined with the (3)H-labeled desmethylimipramine deposition technique. The spatial pattern of flow distribution was quantitated by the coefficient of variation of regional flows (CV, related to global flow heterogeneity) and the correlation between adjacent regional flows (CA, inversely related to local flow randomness). CV was lower in LV than in RV [P < 0.05, nonparametric 2-way analysis of variance (NANOVA)]. When resolution was lowered from 0.1 x 0.1- to 1 x 1-mm(2) pixels, CV decreased by 70% in both LV and RV. CA was higher in LV than in RV (P < 0.05, NANOVA); the interventricular difference in CA was large over the resolutions between 0.4 x 0.4- and 1 x 1-mm(2) pixels. In LV, both CV and CA increased with depth of myocardium (P < 0.05, NANOVA); in subendocardium CV was high comparable with CV in RV (P = 0.47, NANOVA). The enhancement of myocardial workload decreased CV and tended to decrease CA in LV subendocardium (P < 0.05, P = 0.06, respectively; NANOVA). We conclude that 1) microregional flow distribution is less heterogeneous and less random in LV than in RV; 2) transmurally, in LV subendocardium global flow heterogeneity was the highest whereas local flow randomness was the lowest, so that clusters of low- or high-flow regions exist in this LV layer; and 3) global flow heterogeneity decreased and local flow randomness tended to increase (flow homogenizing occurred) in LV subendocardium with increasing myocardial workload. Thus the distributed pattern of myocardial microregional flows may be adaptable to local myocardial metabolic change.