Three genes encoding T-type Ca2+ channels have been described but their correspondence to the various native T-type Ca2+ currents remains uncertain. In particular, Ca(V)3.2 (or alpha1H) was cloned from a human heart library, its message was found abundantly in cardiac tissue, and expressed Ca(V)3.2 was shown to conduct low voltage-activated currents, which inactivate rapidly and are sensitive to Ni2+ and mibefradil. These observations suggested that Ca(V)3.2 might encode native cardiac T-type Ca2+ channels but more information on the pharmacology of Ca(V)3.2 was needed to confirm this hypothesis. In the present study, we compare the pharmacology of Ca(V)3.2 expressed in HEK293 cells and of native T-type Ca2+ channels in guinea pig atrial myocytes ("native-T"). (1) Ca(V)3.2 and native-T are insensitive to TTX and to toxins selective for N-, P-, or Q-type Ca2+ channels (omega-CTx-GVIA, omega-Aga-IVA, omega-CTx-MVIIC). (2) The half-maximal blocking concentration (IC50) of mibefradil on Ca(V)3.2 is near that on native-T and the block is similarly voltage-dependent. (3) Ca(V)3.2 is five- to sixfold less sensitive than native-T to the 1,4-dihydropyridine (DHP) amlodipine, suggesting a difference in the DHP binding site. (4) Both channels display similar (but not identical) sensitivities to the inorganic blockers Ni2+ and Cd2+ and the IC50s are in the range of values found for T-type Ca2+ currents in other cell types. (5) Ni2+ shifts the voltage dependence of Ca(V)3.2 activation but not that of native-T. The many similarities between the two channels support the contention that Ca(V)3.2 encodes cardiac T-type Ca2+ channels. The slight differences may be due to species variations and/or to the choice of splice variant.