Theoretical best performance times (t(theor)) in track running are calculated as follows. Maximal metabolic power ((E)over dot(max)) is a known function of maximal oxygen uptake ((V)over dotO(2max)), of maximal anaerobic capacity (AnS) and of effort duration to exhaustion (t(e)): (E)over dot(max)=f (t(e)). Metabolic power requirement ((E)over dot(r)) to cover the distance (d) in the performance time t(p) is the product of the energy cost of locomotion per unit distance (C) and the speed: (E)over dot(r)=Cxd/t(p). The time values for which (E)over dot(max) (t(e))=(E)over dot(r) (t(p)), assumed to yield t(theor), can be obtained for any given subject and distance provided that (V)over dotO(2max), AnS and C are known, and compared with actual best performances (t(act)). For 15 mingreater than or equal tot(e)greater than or equal to100 s, the overall ratio t(act)/t(theor) was rather close to 1.0. To estimate the relative role of the different factors limiting (V)over dotO(2max), several resistances to O-2 transport are identified, inversely proportional to: alveolar ventilation (R-V*), O-2 transport by the circulation (R-Q), O-2 diffusion from capillary blood to mitochondria (R-t), mitochondrial capacity (R-m). Observed changes of (V)over dotO(2max) are accompanied by measured changes of several resistances. The ratio of each resistance to the overall resistance can therefore be calculated by means of the O-2 conductance equation. In exercise with large muscle groups (two legs), R-Q is the major (75%) limiting factor downstream of the lung, its role being reduced to 50% during exercise with small muscle groups (one leg). R-t and R-m account for the remaining fractions. In normoxia R-V* is negligible; at high altitude it increases progressively, together with R-t and R-m, at the expense of R-Q.