1. We measured the ratio of ETA and ETB sub-types in the media (containing mainly smooth muscle) of human cardiac arteries (aorta, pulmonary and coronary), internal mammary arteries and saphenous veins. 2. In saturation experiments, [125I]-endothelin-1 ([125I]-ET-1) bound with high affinity to the media of each vessel (n = 3 individuals or homogenate preparations +/- s.e. mean): coronary artery, KD = 0.14 +/- 0.02 nM, Bmax = 71.0 +/- 21.0 fmol mg-1 protein; pulmonary artery, KD = 0.85 +/- 0.25 nM, Bmax = 15.2 +/- 10.3 fmol mg-1 protein; aorta, KD = 0.51 +/- 0.02 nM, Bmax = 9.4 +/- 4.4 fmol mg-1 protein; internal mammary artery. KD = 0.34 +/- 0.31 nM, Bmax = 2.0 +/- 0.5 fmol mg-1 protein and saphenous vein, KD = 0.28 +/- 0.05 nM, Bmax = 52.8 +/- 1.0 fmol mg-1 protein. In each vessel, over the concentration-range tested, Hill slopes were close to unity and a one site fit was preferred to a two site model. 3. In competition binding assays, the ETA selective ligand, BQ123 inhibited the binding of 0.1 nM [125I]-ET-1 to the media in a biphasic manner. In each case, a two site fit was preferred to a one or three site model: coronary artery, KDETA = 0.85 +/- 0.03 nM, KDETB = 7.58 +/- 2.27 microM, ratio = 89:11%; pulmonary artery, KDETA = 0.27 +/- 0.05 nM, KDETB = 24.60 +/- 5.34 microM, ratio = 92:8%; aorta, KDETA = 0.80 +/- 0.40 nM, KDETB = 2.67 +/- 2.60 microM ratio = 89:11%; saphenous vein, KDETA = 0.55 +/- 0.17 nM, KDETB = 14.4 +/- 0.26 microM, 85:15% (n = 3 individuals or homogenate preparations +/- s.e. mean). BQ123 showed up to 18000 fold selectivity for the ETA over the ETB sub-type. The ETA-selective ligand, [125I]-PD151242 labelled 85% of the receptors detected by a fixed concentration of [125I]-ET-1 in media of internal mammary artery, measured by quantitative autoradiography. In contrast, the density of ETB receptors detected with [125I]-BQ3020 was 7.0 +/- 1.5 amol mm-2, representing about 8% of [125I]-ET-1. 4. A single band corresponding to the expected position for mRNA encoding the ETA receptor (299 base pairs) was found in the media in each of the five vessels (n = 3 individuals) using reverse transcript as epolymerase chain reaction assays. A single band corresponding to the ETB sub-type (428 base pairs) was also always detected.5. 35S-labelled antisense probes to ETA and ETB hybridised to the media of epicardial coronary arteries as well as intramyocardial vessels, confirming the presence of mRNA encoding both sub-types in the vascular smooth muscle of the vessel wall.6 Although mRNA for both receptors was detected, competition binding using BQ123 demonstrated that the majority (at least 85%) of ET receptors present in smooth muscle are the ETA sub-type. These results provide further support for the hypothesis that the ETA sub-type is the receptor that must be blocked in humans to produce a beneficial vasodilatation in pathophysiological conditions where there is an increase in peptide concentration or receptor density.