1. The influence of the transverse (T) tubules on surface action potential conduction was investigated by comparing electrophysiological and confocal microscopic assessments of tubular changes in osmotically shocked and control fibres from frog sartorius muscle. 2. The membrane-impermeant fluorescent dye, di-8-ANEPPs spread readily from the bathing extracellular solution into the tubular membranes in control, intact fibres. Prior exposure of muscles to a hypertonic glycerol-Ringer solution, its replacement by an isotonic Ca(2+)-Mg(2+) Ringer solution and cooling sharply reduced such access. In contrast, dye application in the course of this osmotic shock procedure stained the large tubular vacuoles hitherto associated with successful muscle detubulation. 3. Conduction velocities in intact, control fibres (1.91 +/- 0.048 m s(-1), mean +/- S.E.M., n = 32 fibres) agreed with earlier values reported at room temperature (18-21 degrees C) and were unaffected by prior episodes of steady cooling to 8-10 degrees C (1.91 +/- 0.043 m s(-1), n = 30). 4. Cooling to 11.5 degrees C reduced these velocities (1.47 +/- 0.081 m s(-1), n = 25) but action potential waveforms still included early overshoots and the delayed after-depolarizations associated with tubular electrical activity. 5. In contrast, action potentials from cooled, superficial fibres in osmotically shocked muscles lacked after-depolarization phases implying tubular detachment. Their mean conduction velocities (1.62 +/- 0.169 m s(-1), n = 25) were not significantly altered from values obtained in untreated controls or in intact fibres in muscle similarly treated with glycerol, in direct contrast to earlier results. 6. Cooling produced similar reductions in maximum rates of voltage change dV/dt in action potentials from all fibre groups with lower rates of change shown by detubulated fibres. 7. Use of an antibody to a conserved epitope of the alpha-subunit of voltage-gated sodium channels suggested a concentration of sodium channels close to the mouths of the T tubules. 8. These electrophysiological and anatomical findings are consistent with a partial independence of electrical events in the transverse tubules from those responsible for the rapid conduction of surface regenerative activity. 9. The findings are discussed in terms of a partial separation of the electrical activity propagated over the surface membrane, from the initiation of propagated activity within the T tubules, by the triggering of the sodium channels clustered selectively around the mouths of the T tubules.