Mitochondrial volume density (Vv [mit]) distributions were measured with a test pattern of concentric rings centered upon randomly chosen capillaries in oxidative skeletal muscle cells of two Antarctic fishes, Trematomus newnesi and Notothenia gibberifrons. Vv(mit) in both species was highest in the ring closest to the capillary, minimal further from the capillary (at a distance that was characteristic for each species), and rose in the annuli furthest from the capillary. Plots of Vv(mit) against total area between each ring and the central capillary fit the form of a second-order polynomial (r2 greater than 0.9). If PO2 or blood-borne metabolite concentration predicates the pattern of Vv(mit) distribution, minimal Vv(mit) is at the same position as the minimum in concentration or gaseous partial pressure of capillary-supplied commodities. This minimum is the boundary between cylinders of tissue being supplied by adjacent capillaries, and thus delineates the maximal diffusion-distance for capillary-supplied commodities. Maximal diffusion-distance (microns) for T. newnesi = 26.23 +/- 1.64; N. gibberifrons = 21.45 +/- 0.51. For O2, maximal diffusion-distance conventionally is referred to as Krogh's radius, R. With an easily obtained estimate of numerical capillary density, these R values can be used to calculate a capillary tortuosity constant (c[k,0]) and capillary length density (Jv[c,f]). c(k,0) values were also determined using an established method, and R and Jv(c,f) values calculated from these values did not significantly differ from values determined from mitochondrial distributions. Mitochondrial distribution analysis may more accurately reflect changes in capillary blood flow and heterogeneity of diffusion and solubility constants within muscle than currently existing techniques. Similar distributions of Vv(mit) reported for several species of vertebrates suggest wide applicability of the method.