The precise mechanisms underlying hypoxic pulmonary vasoconstriction (HPV) are still elusive. The recent discovery of K+ channels that are depressed by hypoxia in pulmonary vascular smooth muscle has provided a potential signal transduction mechanism for linking a reduction in Po2 to Ca2+ entry, but there are many reports suggesting that sustained HPV depends on the presence of the endothelium. Many endothelium-derived vasoactive factors have been investigated as possible mediators of HPV, including endothelium-derived relaxing factor (EDRF-NO), leukotrienes, prostanoids and endothelin, yet none have been found to be indispensable for HPV. They do, however, act as powerful modulators of the response to hypoxia. HPV is probably multifactorial in origin, as exemplified by the biphasic response to hypoxia seen in isolated pulmonary arteries over 40 min. The first phase is of rapid onset but transient, endothelium independent and partly related to Ca2+ release from intracellular stores. The second phase is slowly developing but sustained, dependent on the endothelium and associated with a stable elevation of cytosolic Ca2+. Since tension continues to rise while intracellular [Ca2+] remains constant, this implies Ca2+ sensitization of the contractile apparatus. This is independent of protein kinase C or pH. It is proposed that HPV depends upon both smooth muscle and endothelium resident mechanisms. Inhibition of K+ channels causes an elevation of cytosolic Ca2+, which may not be sufficient to generate substantive contraction on its own. However, release from the endothelium of an as yet unidentified mediator increases Ca2+ sensitivity of the contractile apparatus, and sustained contraction ensues.