1 Intra-arterial injection of endothelin-1 (ET-1, 400pmol; 1 ,ug) in guinea-pig isolated perfused lungs, induced increases in pulmonary inflation pressure (PIP) and perfusion pressure (PPP), associated with oedema formation and thromboxane B2 (TxB2) release but not with the generation of sulphidopeptide leukotrienes or release of histamine. In contrast, aerosol administration of ET-1 (3, 6, lOpgml l, for 2min) evoked a dose-dependent increase in PIP, without significant changes in PPP, oedema formation or TxB2 release. 2 Addition of indomethacin (5pM) or BW 755C (10 or 100pM), but not nordihydroguaiaretic acid (NDGA, 50OuM) or FPL 55712 (10pM), to the perfusion medium led to a significant inhibition of the increases in PIP and PPP, TxB2 release and oedema formation evoked by intra-arterial injection of 400pmol ET-1. In contrast, indomethacin (5pM), BW 755C (100pyM) or FPL 55712 (10UM), added to the perfusion medium 10min prior to challenge, did not affect the increase in PIP induced by a 2-min aerosol of a solution of ET-1 lO jug ml -1. 3 In vivo aerosol administration of indomethacin (100mg ml 1, for 20 min) to non-anaesthetized guineapigs, 15min before lung removal, did not modify the bronchopulmonary response evoked in isolated perfused lungs by an aerosol of ET-1 lOpgmlP . However, under the same experimental conditions, indomethacin significantly inhibited TxB2 release evoked by aerosolized arachidonic acid (2 mg ml-1). 4 In conclusion, the present study shows that when injected by the intra-arterial route, ET-1 effects are mediated primarily via the generation of cyclo-oxygenase metabolites of arachidonic acid, whereas when the aerosol route is used, the peptide appears to act on airway smooth muscle cells, through an indomethacin-insensitive process which may involve some other, as yet unidentified, mediator(s).