Isolated, living muscle fibres from either Xenopus or mouse were observed in a confocal microscope and t-tubules were visualized with sulforhodamine B. Observations were made before and after fatiguing stimulation. In addition, experiments were performed on fibres observed in an ordinary light microscope with dark-field illumination. In Xenopus fibres, recovering after fatigue, t-tubules started to show dilatations 2-5 min post-fatigue. These swellings increased in size over the next 10-20 min to form vacuoles. After 2-3 h of recovery the appearance of the fibres was again normal and force production, which had been markedly depressed 10-40 min post-fatigue, was close to control. Vacuoles were not observed in mouse fibres, fatigued with the same protocol and allowed to recover. In Xenopus fibres, fatigued in normal Ringer solution and allowed to recover in Ringer solution with 30-50 mM L-lactate substituting for chloride (lactate-Ringer), the number and size of vacuoles were markedly reduced. Also, force recovery was significantly faster. Replacement of chloride by methyl sulphate or glucuronate had no effect on vacuolation. Resting Xenopus fibres exposed to 50 mM lactate-Ringer and transferred to normal Ringer solution displayed vacuoles within 5-10 min, but to a smaller extent than after fatigue. Vacuolation was not associated with marked force reduction. Mouse fibres, fatigued in 50 mM lactate-Tyrode (L-lactate substituting for chloride in Tyrode solution) and recovering in normal Tyrode solution, displayed vacuoles for a limited period post-fatigue. Vacuolation had no effect on force production. The results are consistent with the view that lactate, formed during fatigue, is transported into the t-tubules where it attracts water and causes t-tubule swelling and vacuolation. This vacuolation may be counteracted in vivo due to a gradual extracellular accumulation of lactate during fatigue.