An appreciation for the potential of skeletal muscle vascular beds for blood flow (blood flow capacity) is required if one is to understand the limits of the cardiorespiratory system in exercise. To assess this potential, an index of blood flow capacity that can be objectively measured is required. One obvious index would be to measure maximal muscle blood flow (MBF). However, a unique value for maximal MBF cannot be measured, since once maximal vasodilation is attained MBF is a function of perfusion pressure. Another approach would be to measure maximal or peak vascular conductance. However, peak vascular conductance is different among skeletal muscles composed of different fiber types and is a function of perfusion pressure during peak vasodilation within muscle composed of a given fiber type. Also, muscle contraction can increase or decrease blood flow and/or the apparent peak vascular conductance depending on the experimental preparation and the type of muscle contraction. Blood flows and calculated values of conductance appear to be greater during rhythmic contractions (with the appropriate frequency and duration) than observed in resting muscle during what is called "maximal" vasodilation. Moreover, dynamic exercise in conscious subjects produces the greatest skeletal muscle blood flows. The purpose of this review is to consider the interaction of the determinants of muscle blood flow during locomotory exercise. Emphasis is directed toward the hypothesis that the "muscle pump" is an important determinant of perfusion of active skeletal muscle. It is concluded that, during normal dynamic exercise, MBF is determined by skeletal muscle vascular conductance, the perfusion pressure gradient, and the efficacy of the muscle pump.