Channels are water-filled membrane-spanning proteins, which undergo conformational changes as they gate, i.e. open or close. These conformational changes affect both the shape of the channel and the volume of the water-filled pore. We measured the changes in pore volume associated with activation, deactivation, C-type inactivation and recovery in an N-terminal-deleted mutant of the Kv1.4 K+ channel (Kv1.4DeltaN) expressed in Xenopus oocytes. We used giant-patch and cut-open oocyte voltage clamp techniques and applied solutes which are too large to enter the pore mouth to exert osmotic pressure and thus favour smaller pore volume conformations. Applied intracellular osmotic pressure (300 mM sucrose) sped inactivation (time constants (tauinactivation): control, 0.66 +/- 0.09 s; hyperosmotic solution, 0.29 +/- 0.04 s; n = 5, P < 0.01), sped deactivation (taudeactivation: control, 18.8 +/- 0.94 ms; hyperosmotic solution, 8.01 +/- 1.92 ms; n = 5, P < 0.01), and slowed activation (tauactivation: control, 1.04 +/- 0.05 ms; hyperosmotic solution, 1.96 +/- 0.31 ms; n = 5, P < 0.01). These effects were reversible and solute independent. We estimated the pore volume change on inactivation to be about 4500 A3. Osmotic pressure had no effect when applied extracellularly. These data suggest that the intracellular side of the pore closes during C-type inactivation and the volume change is similar to that associated with activation or deactivation. This is also similar to the pore volume estimated from the crystal structure of KcsA and MthK K+ channels. Intracellular osmotic pressure also strongly inhibited re-opening currents associated with recovery from inactivation, which is consistent with a physical similarity between the C-type inactivated and resting closed state.