Local hypoxia dilates systemic and constricts pulmonary blood vessels in mammals without neural or humoral involvement. The direct effects of hypoxia on isolated vessels from bony fish have not been examined. In the present study, isolated vessels (efferent branchial artery, EBA; coeliacomesenteric artery, CMA; ventral aorta, VA; and anterior cardinal vein, ACV) from rainbow or steelhead trout (Oncorhynchus mykiss) were subjected to either passive load (resting tension) or contracted with a ligand or 50 mM KCl and then subjected to 60 min of hypoxia by N(2) administration and an additional 30 min of normoxia. All vessels were usually refractory to hypoxia under conditions of resting tension. EBAs, CMAs and VAs pre-contracted with a receptor-mediated ligand were all significantly relaxed by hypoxia and only VAs recovered significantly upon subsequent restoration of normoxia. In contrast, tension in all arteries pre-contracted with 50 mM KCl was elevated further in response to hypoxia. Conversely, ligand-contracted ACVs responded to hypoxia with a further increase in tension, whereas KCl-contracted ACVs relaxed. During apparently random 2-3-week periods EBA and CMA from steelhead and EBA from rainbow trout were hyper-reactive to hypoxia. Steelhead vessels responded to hypoxia with a rapid contraction that increased in magnitude over 3 days. These contractions were independent of pre-stimulation and they were dose-dependent upon PO(2). In isolated gills, hypoxic perfusate produced an immediate but transient elevation of resistance (R(GILL)) in all four gill arches. R(GILL) increased by as much as 30% of initial values and this response was unaltered upon a second hypoxic exposure. These studies demonstrate that isolated vascular segments of rainbow trout are indeed responsive to hypoxia and that these differential responses are vessel and tone dependent and the overall response may be altered by as yet unknown seasonal or environmental factors. Hypoxia-induced arterial relaxation is blocked by elevated external [K(+)], implicating alteration of transmembrane K(+) conductance and/or membrane potential in this depressor response. K(+)-channel closure or voltage-gated Ca(2+) influx cannot account for arterial vasoconstriction due to hypoxia during KCl contractions. Vascular responses to hypoxia could have a profound impact on local flow in vivo and could mediate ventilation-perfusion matching in the branchial circulation of fish.