Earlier autoradiographic studies with the 5-HT2 receptor agonist [(125)I](±)DOI in human brain showed unexpected biphasic competition curves for various 5-HT2A antagonists. We have performed similar studies in rat brain regions with selective 5-HT2A (M100907) and 5-HT2C (SB242084) antagonists together with ketanserin and mesulergine. The effect of GTP analogues on antagonist competition was also studied. Increasing concentrations of Gpp(NH)p or GTPγS resulted in a maximal inhibition of [(125)I](±)DOI-specific binding of approximately 50 %. M100907 competed biphasically in all regions. In the presence of 100 μM Gpp(NH)p, M100907 still displaced biphasically the remaining [(125)I](±)DOI binding. Ketanserin showed biphasic curves in some regions and monophasic curves in others. In the latter, Gpp(NH)p evidenced an additional high-affinity site. SB242084 competed biphasically in brainstem nuclei and monophasically in the other regions. In most areas, SB242084 affinities were not notably altered by Gpp(NH)p. Mesulergine competed monophasically in all regions without alteration by Gpp(NH)p. These results conform with the extended ternary complex model of receptor action: receptor exists as an equilibrium of multiple conformations, i.e. ground (R), partly activated (R*) and activated G-protein-coupled (R*G) conformation/s. Thus, [(125)I](±)DOI would label multiple conformations of both 5-HT2A and 5-HT2C receptors in rat brain, and M100907 and ketanserin would recognise these conformations with different affinities.