The exposure of amphibian muscle to osmotic shock through the introduction and subsequent withdrawal of extracellular glycerol causes 'vacuolation' in the transverse tubules. Such manoeuvres can also electrically isolate the transverse tubules from the surface ('detubulation'), particular if followed by exposures to high extracellular [Ca2+] and/or gradual cooling. This study explored factors influencing vacuolation in Rana temporaria sartorius muscle. Vacuole formation was detected using phase contrast microscopy and through the trapping or otherwise of lissamine rhodamine dye fluorescence within such vacuoles. The preparations were also examined using electron microscopy, for penetration into the transverse tubules and tubular vacuoles of extracellular horseradish peroxidase introduced following the osmotic procedures. These comparisons distinguished for he first time two types of vacuole, 'open' and 'closed', whose lumina were respectively continuous with or detached from the remaining extracellular space. The vacuoles formed closed to and between the Z-lines, but subsequently elongated along the longitudinal axis of the muscle fibres. This suggested an involvement of tubular membrane material; the latter appeared particularly concentrated around such Z-lines in the electron-micrograph stereopairs of thick longitudinal sections. 'Open' vacuoles formed following osmotic shock produced by extracellular glycerol withdrawal from a glycerol-loaded fibre at a stage when one would expect a net water entry to the intracellular space. This suggests that vacuole formation requires active fluid transport into the tubular lumina in response to fibre swelling. 'Closed' vacuoles only formed when the muscle was subsequently exposed to high extracellular [Ca/+] and/or gradual cooling following the initial osmotic shock. Their densities were similar to those shown by 'open' vacuoles in preparations not so treated, suggesting that both vacuole types resulted from a single process initiated by glycerol withdrawal. However, vacuole 'closure' took place well after formation of 'open' vacuoles, over 25 min after glycerol withdrawal. Its time course closely paralleled the development of detubulation reported recently. It was irreversible, in contrast to the reversibility of 'open' vacuole formation. These findings identify electrophysiological 'detubulation' of striated muscle with 'closure' of initially 'open' vacuoles. The reversible formation of open vacuoles is compatible with some normal membrane responses to some physiological stresses such as fatigue, whereas irreversible formation of closed vacuoles might only be expected in pathological situations as in dystrophic muscle.