In the rabbit, 5,6-epoxyeicosatrienoic acid (EET) was reported both to dilate and to constrict pulmonary blood vessels. We propose that these seemingly contradictory results could be explained by differences in responses to 5,6-EET in large-conductance pulmonary arteries (PA) compared with smaller PA and resistance vessels. Thus we found that in rings of extralobar PA [>2-mm outside diameter (OD)], in which active tension had been increased with PGF₂, 5,6-EET produced relaxation in a concentration- and cyclooxygenase (COX)-dependent manner. In contrast, 5,6-EET increased tension in intralobar (1- to 2-mm OD) PA. Small extralobar PA (2- to 2.5-mm OD) exhibited intermediate responses. In the intact lung, the net effect of 5,6-EET (1 × 10⁸-1 × 10⁵ M) was an increase in pulmonary vascular resistance (PVR) from 13.0 ± 0.5 to 47.8 ± 4.6 mmHg · 100 ml¹ · min¹ (EC₅₀ 5.9 ± 1.7 × 10⁷ M). The increase in PVR was accompanied by a 10-fold increase in perfusate thromboxane (TX)B₂ concentration. The 5,6-EET-induced increase in PVR was prevented with indomethacin (100 µM), a cyclooxygenase inhibitor, or ONO-3708 (20 µM), a TX/PGH₂ (TP) receptor antagonist, but not with OKY-046 (700 µM), a TX synthase inhibitor. These results demonstrate that although 5,6-EET dilates large extralobar PA segments in a COX-dependent manner, in the intact rabbit lung 5,6-EET produces constriction that requires synthesis of a COX-dependent agonist of the TP receptor other than TX.