Alpha-2 adrenoceptors were characterized in three tissue culture cell lines and in membrane homogenates of guinea pig lung via the ability of a series of alpha adrenoceptor antagonists to inhibit the binding of [3H]clonidine, [3H]UK-14,304 and [3H]rauwolscine. The cells studied included those known to possess receptors of alpha-2A (HT29) and alpha-2B (NG108-15) subtypes as well as the previously uncharacterized NCB20 cells. Correlation of the ability of the antagonists to inhibit [3H]clonidine or [3H]UK-14,304 binding did not identify alpha-2 adrenoceptor subtypes. On the other hand, correlation of antagonist affinities against [3H]rauwolscine binding showed HT29 cells and guinea pig lung to have similar characteristics (r = 0.911) as did NG108-15 and NCB20 cells (r = 0.985). These data suggest subtle differences in the binding of [3H]agonists and [3H]antagonists to the alpha-2 adrenoceptor, resulting in the failure of [3H]clonidine and [3H]UK-14,304 to recognize differences between alpha-2A and alpha-2B receptor subtypes. 6-chloro-9-[(3-methyl-2-butenyl)oxy]-3-methyl-1H-2,3,4,5-tetrahydro-3- benzazepine (SK&F 104078) did not differentiate between the alpha-2A and alpha-2B receptor subtypes. However, 2-vinyl-7-chloro-3,4,5,6-tetrahydro-4-methylthieno[(4,3,2ef] benzazepine (SK&F 104856), which has similar selectivity in functional in vitro models, was about 35-fold more potent in displacing [3H]rauwolscine binding to the alpha-2B site. These data provide additional evidence that the functional subclassification of alpha-2 adrenoceptors based on sensitivity to SK&F 104078 and SK&F 104856 subdivides this receptor in a different manner than does the alpha-2A/alpha-2B subclassification scheme.