1. The coiled reabsorptive segment of human sweat ducts was cultured in vitro. Cells were then harvested and plated onto a dialysis membrane which was glued over a hole in a small disc. Cultures were maintained in a low serum, hormone-supplemented medium that allowed the cells to grow to confluency. The disc was then placed as a partition between two compartments of a miniature Ussing chamber. The chamber was mounted on the stage of an inverted microscope and intracellular potentials were recorded under transepithelial open-circuit or voltage clamp conditions. All values are given as means +/- S.E.M. and n refers to the number of preparations or duct cells. 2. Under control conditions, the cultured epithelia developed mucosa-negative transepithelial potentials (Vt) ranging from -2.5 to -38 mV (-13.5 +/- 1.5 mV, n = 36). The basolateral membrane potential (Vb) was -39.4 +/- 0.7 mV (n = 50 cells), and the apical membrane potential (Va) was linearly correlated with Vt:Va = 1.0 Vt -39.3 mV (r = -0.78, n = 50). 3. The epithelium generated inwardly directed short-circuit currents (Isc) of 12-95 microA cm-2 (45 +/- 4 microA cm-2, n = 36) with a steady-state intracellular potential. Vc = -31.1 +/- 0.6 mV and a fractional resistance of the apical membrane, fR = 0.59 +/- 0.01 (n = 115 cells). 4. The Na+ channel blocker amiloride (mucosal bath, 10 microM) abolished Isc -0.8 +/- 0.6 microA cm-2), the cells hyperpolarized to -61.0 +/- 1.2 mV, and fR increased to 0.85 +/- 0.01 (n = 44). These effects were fully reversible. 5. During initial stimulation with the cholinergic agonist, methacholine (serosa, 5 or 10 microM), the short-circuit current increased to 80 +/- 10 microA cm-2, the cells hyperpolarized to -55.8 +/- 1.2 mV, and fR increased to 0.82 +/- 0.01 (n = 35). 6. In short-circuited preparations stimulated with methacholine an increase in mucosal potassium concentration ([K+]m) from 5 to 25 mM had no significant effect, while a similar increase in the serosal K+ concentration ([K+]s) produced a change in Vc of 44 +/- 3 mV per log10[K+]s (n = 9). In non-stimulated preparations this change was only 16 +/- 2 mV per log10[K+]s (n = 13). After blocking the apical Na+ channels with amiloride the slope was 24 +/- 5 mV per log10[K+]s in unstimulated preparations.(ABSTRACT TRUNCATED AT 400 WORDS)